Size Reduction of Ammonia Scrubbers for Pig and …prairieswine.com/pdf/3365.pdfammonia scrubbers that are applied at animal houses at all times treat the entire exhaust airflow and

ARTICLE IN PRESS

doi:10.1016/j.biosystemseng.2006.05.006SE—Structures and Environment

Biosystems Engineering (2006) 95 (1), 69–82

Size Reduction of Ammonia Scrubbers for Pig and Poultry Houses: Use ofConditional Bypass Vent at High Air Loading Rates

R.W. Melse; A.V. van Wagenberg; J. Mosquera

Wageningen University and Research centre (WUR), Animal Sciences Group, PO Box 65, 8200 AB Lelystad, Netherlands;e-mail of corresponding author: [email protected]

(Received 21 June 2005; accepted in revised form 9 May 2006; published online 24 July 2006)

In The Netherlands, both acid and biological air scrubbers are used for removal of ammonia from exhaust airat pig and poultry houses. Current regulations require that scrubbers are dimensioned for treating themaximum airflow rate that may occur, so on average these systems are overdimensioned and underloaded. Anew approach is introduced that is based on bypassing airflow peaks untreated. As a result, the air loading ratein m3 [air] m�3 [scrubber] h�1 and ammonia loading rate in kg [NH3] m

�3 [scrubber] h�1 of the scrubber aremore constant in time and average loading rates increase. By model calculations and analyses of measurementdatasets it was demonstrated that the application of such a scrubber significantly decreases the requiredscrubber size while ammonia emission levels are only slightly increased (e.g. where the bypass is operated at50% of the maximum ventilation rate and the scrubber volume is reduced by 50%, the bypass venting systemsonly allows 10–20% of the total ammonia load to be vented untreated). As a result, both the efficiency ofscrubber utilisation in kg [NH3 removal] m�3 [scrubber volume] and the cost-effectiveness of air scrubbing forammonia removal in kg [NH3 removal] h�1 are increased.r 2006 IAgrE. All rights reserved

Published by Elsevier Ltd

1. Introduction

1.1. Animal husbandry and ammonia abatement

Pig and poultry production contributes substantiallyto the economies of many Western European countriesin terms of employment and export of products. Pigproduction in Western Europe is concentrated in severalregions characterised by large-scale intensive farms. TheNetherlands, with 16 million inhabitants and a popula-tion density of about 400 inhabitants per km2, houses 11million pigs at approximately 10 000 farms (CBS, 2004).From the 1980s onwards, the emission of ammonia

(NH3) from livestock farming has become a majorenvironmental concern because ammonia emission isone of the three main sources of soil acidification andeutrophication of natural soils in The Netherlands (Heij& Erisman, 1995, 1997). Therefore considerable effortshave been put into the development of ammoniaabatement techniques in animal operations. This focus

1537-5110/$32.00 69

on ammonia abatement has resulted in the developmentof a variety of low-emission livestock housing systemsthat are applied today. These systems include end-of-pipe systems for treatment of the exhaust air from pigand poultry houses, viz. acid scrubbers and bioscrubbersor biotrickling filters. However, investment and opera-tional costs of scrubber systems are generally consideredas high.

1.2. Working principle of ammonia scrubbers

An air scrubber, or trickling filter, is a reactor that hasbeen packed with an inert packing material. Thepacking material usually has a large porosity, or voidvolume, and a large specific area. The packed bed iswetted by spraying water on top. Exhaust air from ananimal house, containing ammonia, is introduced eitherin a cross-current or counter-current direction whichresults in intensive contact between air and water

r 2006 IAgrE. All rights reserved

Published by Elsevier Ltd

ARTICLE IN PRESS

R.W. MELSE ET AL.70

enabling mass transfer from air to liquid phase. Aschematic of a trickling filter is given in Fig. 1.

In an acid scrubber, the pH of the recirculation wateris kept below 4 by the addition of acid, usually sulphuricacid. The ammonia dissolves in the liquid phase and iscaptured by the acid forming an ammonium saltsolution which is discharged on a regular basis andreplaced with fresh water. The average ammoniaremoval efficiencies of acid scrubbers may vary from40% to 100% (overall average of 96%) as wasdemonstrated in on-farm research at three pig and twopoultry sites (Melse & Ogink, 2005).

In a biotrickling filter, bacteria are present partly as abiofilm on the packing material and partly suspended inthe water that is being recirculated. The dissolvedammonia is converted into nitrite (NO2

�) and subsequentlyinto nitrate (NO3

�) by nitrification which is mainly carriedout by Nitrosomonas and Nitrobacter species, respectively(Focht & Verstraete, 1977; Prosser, 1986). The liquidphase, containing nitrite and nitrate, is discharged on aregular basis and replaced with fresh water. The averageammonia removal efficiencies of biotrickling filters mayvary from negative values to 100% (overall average of70%) as was demonstrated in on-farm research at five pigsites, one poultry, and one veal calve site (Melse & Mol,2004; Melse & Ogink, 2005).

In The Netherlands about 160 acid scrubbers and 45biotrickling filters are in operation for ammoniaremoval from ventilation air of animal houses, mainlyat pig and poultry farms (Melse & Ogink, 2005; Melse &Willers, 2005).

1.3. Air scrubber dimensioning

For fattening pigs, ventilation rates vary from 10 to90m3 animal�1 h�1, depending on animal production

Air outlet Fresh water supply

RecirculationPacking

Water dischargeAir inlet Buffertank

Fig. 1. Schematic of an air scrubber; from Melse and Ogink(2005)

stage and weather conditions (Ogink & Lens, 2001;Seedorf et al., 1998). The average year-round ventilationrate for fattening pigs is about 35m3 animal�1 h�1 atDutch weather conditions, i.e. a moderate maritimeclimate. For broilers, ventilation rates usually vary from1 to 3�6m3 kg�1 [liveweight] h�1 (ASG, 2004; Seedorf et

al., 1998), which equals 0�04–9m3 broiler�1 h�1, depend-ing on weather conditions. The average year-roundventilation rate for broilers is about 3m3 animal�1 h�1.Minimum empty bed air residence times (EBRT) ofammonia scrubbers, i.e. scrubber volume in m3 dividedby maximum airflow rate in m3 s�1, have typically valuesof 0�5–1�0 s, which equals air loading rates from 7200 to3600m3m�3 [scrubber volume] h�1.

Pig houses with separated room ventilation systemsare usually equipped with a central ventilation system ifan air scrubber system is applied. In a central ventilationsystem, one or more central fans withdraw air from therooms into a central ventilation duct, so air that exits thecentral ventilation duct and subsequently enters thescrubber system is a mixture of the exhaust air from allrooms. The airflow rate and ammonia concentrationvary from room to room if in each room a batch isstarted at a different date. The use of a centralventilation system, which mixes the exhaust air fromseveral rooms that each contain pigs of a different age,will reduce these variations and result in a moreconstant loading rate of both air in m3 m�3 [scrubbervolume] h�1 and ammonia in kg [NH3] m

�3 [scrubbervolume] h�1. The more constant air loading rate resultsin a lower average ventilation rate per animal place, thusthe installed ventilation capacity will be lower than for aventilation system per room. Consequently, the requiredscrubber size for treating all air will be lower. Theforegoing is not the case for a pig house in which allrooms are operated simultaneously and where all pigs inthe house have the same age.

1.4. Air scrubber with bypass vent

Although air scrubbing systems for ammonia removalfrom animal house exhaust air are commerciallyavailable in The Netherlands and considered as off-the-shelf techniques today, investment and operationalcosts are relatively high as compared to other availablehousing systems with emission reduction techniques.Current regulations in The Netherlands require thatammonia scrubbers that are applied at animal houses atall times treat the entire exhaust airflow and meet therequired minimum removal efficiency. Therefore con-ventional scrubbers are designed for treating themaximum exhaust airflow rate, even while this max-imum airflow rate only occurs for a short period of time.

ARTICLE IN PRESS

SIZE REDUCTION OF AMMONIA SCRUBBERS 71

So most of the time these scrubbers are oversized andunderloaded, even if they are connected to a centralventilation system.However, a significant reduction of the investment

and operational costs of air scrubbing is achieved ifrequired scrubber volumes can be decreased so thatscrubbers are dimensioned for lower maximum ventila-tion rates. This could be achieved by combining an airscrubber with an air bypass system that conditionallyallows a part of the exhaust air to bypass the scrubberand be vented untreated. In this way a scrubber can bedesigned for the average airflow as peaks in the airflowwill be bypassed. In such a system, the bypass vent is setto operate whenever the actual airflow rate exceeds themaximum airflow rate the scrubber has been designedfor, i.e. the bypass setpoint airflow rate. Only the part ofthe air that exceeds this setpoint is bypassed, so thescrubber is operated just at its designed airflow rate. It isnot clear beforehand to what extent the ammoniaemission rate will be increased if part of the exhaustair is vented untreated, as the ammonia concentration ofthe air varies over time.

1.5. Objectives and approach

The first aim of this paper is to analyse howventilation rate and ammonia emission from pig andpoultry houses are interrelated and how the conditionalbypassing of part of the exhaust air of the scrubbersystem affects the ammonia emission. The second aim isto determine how conditional bypassing affects therequired scrubber size, because a reduction of scrubbersize will increase the economic feasibility of ammoniaremoval by air scrubbing.In order to achieve these aims, previously gathered

experimental datasets of ammonia emission and ventila-tion rate during batches of broiler and fattening pigs(Table 1) are analysed. For fattening pigs, a model isdeveloped for year-round simulations of the ammoniaemission when using a scrubber with a bypass vent, bothfor a central ventilation system and for a ventilationsystem per room.

2. Materials and methods

2.1. Ammonia emission datasets for fattening pigs and

broilers

From previously carried out research, six datasetswere analysed that contained continuous measurementsof the ventilation rate and ammonia concentration ofexhaust air from fattening pigs (dataset 1–4) and broilers

(dataset 5 and 6) in conventional housing systems. Allhouses were equipped with a mechanical ventilationsystem that is controlled by the temperature of the roomair; ventilation and temperature setpoints, heat produc-tion by the animals and outside conditions vary in timeduring a batch, resulting in a varying ventilation rate. InTable 1 some details are given for these datasets.

For each dataset, the ventilation rate in m3 h�1 wasmeasured continuously by a calibrated ventilation ratesensor in the ventilation shaft. The exhaust air wascontinuously sampled using a vacuum pump at a fixedflow rate which was controlled by a critical orifice(0�5 lmin�1) and led to a NH3 converter/NOx-analysersystem in which determination of the ammonia con-centration in mg m�3 took place (Mosquera et al., 2002).All sampling tubes had been made of Teflon, wereisolated, and heated with a coil of resistance wire to atemperature of approximately 20 1C above the ambienttemperature to prevent condensation of water andsubsequent adsorption of ammonia. The data of thecontinuously measured ventilation rate and ammoniaconcentration were averaged over 1-h intervals and theresulting datasets were analysed in this study.

All datasets were gathered during rather short periods(several months) with the specific weather conditions atthat time and place. Furthermore, each dataset forfattening pigs contains the measurements of the ventila-tion rate and ammonia concentration of a single pigroom. In practice, however, scrubber systems are usuallyinstalled at pig houses that are equipped with a centralventilation system.

In order to estimate the effect of the use of a scrubbersystem with a bypass vent on the year-round emission ofammonia from a fattening pig house without thesedrawbacks, both for a central ventilation system and fora ventilation system per room, model calculations weredone of ventilation rate and ammonia emission.

2.2. Modelling of the ventilation rate of a pig house

Model calculations of ventilation rates and ammoniaemission were done to describe the year-round emissionpattern of a fattening pig house. This pattern was usedto estimate the effect of the use of a scrubber systemwith a bypass vent on the ammonia emission.

For calculating the ventilation rate as a function ofoutside temperature and animal production stages (dayof batch), relationships were used that had beencalculated by Van Wagenberg and Vermeij (2001a,2001b) with the simulation model ANIPRO (VanOuwerkerk, 1999); these relationships are shown inFig. 2. The ANIPRO model has been based on earlierdetailed studies on heat production of pigs (Bruce &

ARTICLE IN PRESS

Table

1Datasets

ofhourlyaveraged

ventilationrate

andammonia

emissionmeasuredatconventionalhousingsystem

sforpigsandpoultry�

Da

tase

tn

o.

So

urc

eM

easu

rin

gp

erio

dN

o.

of

an

ima

lsp

rese

nt

Ro

om

tem

per

atu

re,

min

–m

ax

(a

vera

ge)

,1C

Ven

tila

tio

nra

te,

min

–m

ax

(a

vera

ge)

,m

3

an

ima

l�1

h�

1

Am

mo

nia

emis

sio

n,

min

–m

ax

(a

vera

ge)

,m

ga

nim

al�

1h�

1

1Satter

eta

l.(1997)

Ba

tch

1:

13March–9July

1996

130fatteningpigs

19�1–34�9

(23�1)

5�3–57�5

(28�8)

45–640(218)

Ba

tch

2:

17July–18Novem

ber

1996

130fatteningpigs

19�2–31�0

(22�9)

7�5–53�7

(27�6)

108–514(277)

2Groenestein

andHuisin

‘tVeld(1996)

15July–9Novem

ber

1995

110fatteningpigs

17�6–31�4

(22�7)

8�5–65�0

(42�9)

32–672(330)

3HolandGroenestein

(2005)

Ba

tch

1:

4June–10October

2002

80fatteningpigs

19�3–33�1

(24�3)

22�9–66�8

(49�6)

155–551(402)

Ba

tch

2:

23October

2002–16

January

2003

80fatteningpigs

17�7–25�6

(20�7)

14�6–43�6

(24�0)

180–628(393)

4Huisin

‘tVeldand

Groenestein

(1995)

1June–28September

1994

64fatteningpigs

16�1–33�5

(23�3)

11�6–112�4

(65�8)

62–562(252)

5Wever

eta

l.(1999)

Ba

tch

1:

22July–31August

1998

41040broilers

15�9–33�7

(26�5)

0�13–3�44(1�67)

0�0–42�7

(11�9)

Ba

tch

2:

16October–23Novem

ber

1998

40630broilers

12�1–34�0

(26�5)

0�08–1�99(0�79)

0�0–41�0

(8�8)

6HolandGroot

Koerkamp(1998)

Ba

tch

1:

11July–20August

1997

11925broilers

21�2–32�8

(26�7)

0�14–4�82(1�80)

0�1–12�7

(4�1)

Ba

tch

2:

2September–13October

1997

10900broilers

18�3–36�7

(25�4)

0�10–5�25(1�48)

0�1–28�5

(7�1)

Ba

tch

3:

25October–5Decem

ber

1997

11000broilers

18�7–33�9

(25�0)

0�10–2�06(0�88)

0�1–12�8

(5�8)

Ba

tch

4:

24July–2September

1998

10865broilers

10�6–32�9

(25�7)

0�10–4�35(1�64)

0�1–65�6

(15�5)

�Allfarm

swerelocatedin

TheNetherlands.

R.W. MELSE ET AL.72

ARTICLE IN PRESS

SIZE REDUCTION OF AMMONIA SCRUBBERS 73

Clark, 1979; Sterrenburg & Van Ouwerkerk, 1986a,1986b). The conditions used as input for the ANIPROmodel were based on a conventional situation that isrepresentative for pig houses in The Netherlands, i.e.

mechanically ventilated and insulated. It is assumed thatthe fattening pigs have a weight of 23 kg at the start and110 kg at the end of a batch of 110 days.For the outside air temperature, i.e. the temperature

of the inlet air of the ventilation system, a meteorolo-gical reference year was used which is representative forDutch weather conditions and contains year-roundhourly temperature values (Lund, 1984). The tempera-ture profile of this reference year is shown in Fig. 3.Simulations of the ventilation rate were done both for

a central ventilation system and for a ventilation systemper room. For the pig house with the central ventilationsystem, it is assumed that the house contains 12 roomswhich are operated non-simultaneously, i.e. the animalsdiffer 9 days in age from room to room, so every 9 daysa new batch is started in one of the rooms.

2.3. Modelling of the ammonia emission from a pig house

The ammonia emission rate in kg [NH3] h�1 of pig

houses is correlated to the animal production stage asthe ammonia emission rate depends on the amount ofmanure that has been accumulated inside the animalhouse (Ni et al., 2000; Mosquera et al., 2005). Mosquera

0

10

−15 −10 − 5 0 5

20

30

40

50

60

70

80

90

Outside t

Ven

tilat

ion

rate

, m3

(fat

teni

ng p

ig)−1

h−1

day 1− 7day 7 − 14

day 14 − 28

day 28 − 42

day 42 − 56

day 56 − 70

day 70 − 84

day 84 − 98

day 98 − 112

day 112 − 119

Fig. 2. Ventilation rate for fattening pigs, depending on producWagenberg and V

et al. (2005) analysed 34 ammonia emission datasets thatwere gathered at 19 different pig farm locations in TheNetherlands. The datasets contained hourly-averagedvalues of continuous measurements of the ventilationrate and ammonia concentration of exhaust air fromfattening pigs in both conventional housing systems andhousing systems with emission reduction techniques, viz.manure flushing, manure cooling, or reduction ofemitting manure surface. The average number ofanimals included in the datasets was 95 fattening pigsand the average duration of the measurement periodwas 91 days.

Ammonia emission, calculated as the average of alldatasets, linearly increases during a batch (Fig. 4). Theequation of Fig. 4 was used to model the ammoniaemission year-round as Mosquera et al. (2005) could notdistinguish a seasonal effect on the ammonia emissionpattern.

3. Results and discussion

3.1. Analysis of ammonia emission datasets for fattening

pigs and broilers

3.1.1. Fattening pigs

The ventilation rate varies in time as it is controlled bythe temperature of the exhaust air which in turn dependson pig production stage and outside temperature. For

10 15 20 25 30 35

emperature, °C

tion stage (day of batch) and outside temperature; from Vanermeij (2001a)

ARTICLE IN PRESS

− 10

− 5

0

1 Jan 1 Feb 4 Mar 4 Apr 5 May 5 Jun 6 Jul 6 Aug 6 Sep 7 Oct 7 Nov 8 Dec

5

10

15

20

25

30

35

Out

side

tem

pera

ture

, °C

Date

Fig. 3. Temperature pattern of a meteorological reference year for The Netherlands (hourly averages); from Lund (1984)

y = 1.0 x + 46.7R 2 = 1.0

Am

mon

ia e

mis

sion

leve

l, %

of

aver

age

emis

sion

160

140

120

100

80

60

40

20

00 20 40 60 80 100

Day of batch

Fig. 4. Ammonia emission pattern of fattening pigs during a batch (day 1–105), expressed as percentage of the average emissionlevel (average of 34 emission datasets); the equation of the trendline and the coefficient of determination are given in the text box;

from Mosquera et al. (2005); R2, coefficient of determination

R.W. MELSE ET AL.74

datasets 1–4 (pig houses), the cumulative frequencydistribution of the ventilation rate was determined, i.e.

for every value of the ventilation rate it was calculatedwhat percentage of the time the actual air flow had beenbelow this value. In Fig. 5, as an example, thisrelationship is plotted for dataset 2. From Fig. 5 it canbe seen that the maximum ventilation rate was65m3 (fattening pig)�1 h�1 [for 100% of the time theventilation rate was below 65m3 (fattening pig)�1 h�1]

and that, for example, for 50% of the time the actualventilation rate was below 45m3 (fattening pig)�1 h�1.

If a scrubber with a bypass system had been installed,from the cumulative frequency distribution pattern itfollows what percentage of the time a part of the airwould be bypassed and vented untreated for a particularbypass setpoint airflow rate. There is a large variation inthe ammonia concentration of the exhaust air fordataset 2 (Fig. 6).

ARTICLE IN PRESS

00 10 20 30 40 50 60 70

10

20

30

40

50

60

70

80

90

100

Ventilation rate, m3 (fattening pig)−1 h−1

Cum

ulat

ive

freq

uenc

y di

stri

butio

n, %

Fig 5. Cumulative frequency distribution of ventilation rate (single room with 110 fattening pigs, dataset 2)

0

5

10

15

20

25

Am

mon

ia c

once

ntra

tion,

mg

m−3

0 10 20 30 40 50 60 70

Ventilation rate, m3 (fattening pig)−1 h−1

Fig. 6. Plot of measured ammonia concentration versus ventilation rate (single room with 110 fattening pigs, dataset 2)

SIZE REDUCTION OF AMMONIA SCRUBBERS 75

With regard to the unwanted emission of ammonia,however, it is necessary to know the amount ofammonia in kg h�1 that is emitted through the bypass.The amount of ammonia that would be venteduntreated if a scrubber system with a bypass would beused, was calculated by combining the hourly measuredventilation rates and ammonia concentrations. In Fig. 7

the bypassed ammonia load is plotted as a function ofthe bypass setpoint flow rate for all datasets. For dataset1, 2, and 3 the trend of Fig. 7 is similar and all have a

maximum ventilation rate of about 60m3 (fatteningpig)�1 h�1, which is a current design airflow rate formechanical ventilation systems in pig houses. Dataset 4has a maximum ventilation rate of over 100m3 (fatten-ing pig)�1 h�1 and is considered as a non-representativesituation.

The average of dataset 1–3 clearly shows that in caseof a bypass setpoint flow rate of 30m3 (fatteningpig)�1 h�1, which is 50% of the maximum airflow rateof about 60m3 (fattening pig)�1 h�1, still 80% of the

ARTICLE IN PRESS

00

10

10 20 30 40 50 60 70 80 90 100

20

30

40

50

60

70

80

90

100

Byp

asse

d am

mon

ia lo

ad,%

of

tota

l loa

d

Dataset 3

Dataset 1

Byepass setpoint, m3 (fattening pig)−1 h−1

Dataset 2

Average of dataset 1− 3

Dataset 4

Fig. 7. Amount of ammonia that would be bypassed if a scrubber with a bypass vent had been installed as a function of the bypasssetpoint (fattening pigs, dataset 1–4)

R.W. MELSE ET AL.76

ammonia load enters the scrubber and only 20% isvented untreated.

Assuming that the maximum air loading rate of thescrubber in m3m�3 [scrubber volume] h�1 remainsunchanged, the size of the scrubber with the bypasssetpoint of 50% is only 50% of the size of aconventional ammonia scrubber without a bypass vent.Assuming the same relative ammonia removal in % asfor a conventional scrubber, the efficiency of scrubberutilisation, expressed as kg [NH3 removal] m�3 [scrub-ber volume], thus increases. The slopes of the tangentlines that can be drawn for the curves of Fig. 7 representthis efficiency. For the curve that indicates the averageof dataset 1–3, starting at the right end of the X axis, thetangent line slope continually decreases (gets morenegative) with decreasing value of x down to a bypasssetpoint of 20m3 (fattening pig)�1 h�1. This means theefficiency of scrubber utilisation has its maximum valueat a bypass setpoint of 20m3 (fattening pig)�1 h�1 orlower. For dataset 4, a similar relationship exists.

3.1.2. Broilers

For the two datasets on ventilation rates andammonia concentrations of broilers, similar calculationswere done as for fattening pigs. In Fig. 8 the ammoniaload is plotted that would be bypassed if a scrubber witha bypass vent had been installed as a function of thebypass setpoint flow rate. As for Fig. 7, starting at theright end of the X axis, the tangent line slopes for thecurves in Fig. 8 continually decrease (get more negative)with a decreasing bypass setpoint flow rate. A decreas-ing tangent line slope means that the efficiency of

scrubber utilisation in kg [NH3 removal] m�3 [scrubbervolume] increases. The average of dataset 5 and 6 clearlyshows that in case of a bypass setpoint flow rate of1�75m3 broiler�1 h�1, which is 50% of the maximumairflow rate of about 3�5m3 broiler�1 h�1, still 85% ofthe ammonia load enters the scrubber and only 15% isvented untreated.

3.2. Model results of the ammonia emission from a

scrubber with a bypass vent at a pig house

3.2.1. Ventilation rate

The ventilation rate was calculated for an individualpig room during 1 year. Every 112 days a new batch isstarted with a low ventilation rate at the start and a highventilation rate at the end date (between batches, roomsare empty for 2 days); the maximum ventilation rate is80m3 (fattening pig)�1 h�1 (Fig. 9).

In Fig. 10 the calculated ventilation rate is shown for acentral ventilation system with 12 rooms during 1 year.The ventilation rate is significantly higher in summerthan in winter. The variation of the average ventilationrate for the whole building (Fig. 10) is much smallerthan for the ventilation rate of a single room (Fig. 9) andbatch start and end dates cannot be distinguishedanymore; the maximum of the average ventilation ratefor the whole building is 62m3 (fattening pig)�1 h�1.The average ventilation rate for both the simulation of asingle room and the simulation of 12 rooms, is 27m3

(fattening pig)�1 h�1.

ARTICLE IN PRESS

0

10

0 0.5 1.5 2.5 3.5 4.5 53 421

20

30

40

50

60

70

80

90

100

Bypass setpoint, m3 broiler −1 h−1

Byp

asse

d am

mon

ia lo

ad, %

of

tota

l am

mon

ia lo

ad

Dataset 5

Dataset 6

Average of dataset 5 and 6

Fig. 8. Amount of ammonia that would be bypassed if a scrubber with a bypass vent had been installed as a function of the bypasssetpoint (broilers, dataset 5 and 6)

0

10

20

30

40

50

60

70

80

Date

Ven

tilat

ion

rate

, m3

(fat

teni

ng p

ig)−1

h−1

1 Jan 1 Feb 4 Mar 4 Apr 5 May 5 Jun 6 Jul 6 Aug 6 Sep 7 Oct 7 Nov 8 Dec

Fig. 9. Model calculation of the ventilation rate of one fattening pig room for a meteorological reference year (hourly averages);every 112 days a new batch is started

SIZE REDUCTION OF AMMONIA SCRUBBERS 77

3.2.2. Ammonia emission

From 13 datasets containing continuous measurementof ventilation rate and ammonia concentrations, eachgathered during several months at nine differentlocations in The Netherlands, Mosquera et al. (2005)calculated that the average ammonia emission fromfattening pigs in a conventional housing system was2�9 kg animal�1 yr�1. Using this value and Fig. 4, theyear-round ammonia emission was calculated for

individual fattening pig rooms. In Fig. 11, the ammoniaemission that was calculated is shown for one of therooms as an example; for every next room, the emissionpattern is shifted by 9 days.

The calculated year-round ammonia emission, asshown in Fig. 11 for a single room, is added for all 12rooms and divided by the total ventilation rate of the pighouse that was modelled (Fig. 10). This results in asimulation of the year-round ammonia concentration of

ARTICLE IN PRESS

0

10

20

30

40

50

60

70

80

Ven

tilat

ion

rate

, m3

(fat

teni

ng p

ig)−1

h−1

1 Jan 1 Feb 4 Mar 4 Apr 5 May 5 Jun 6 Jul 6 Aug 6 Sep 7 Oct 7 Nov 8 Dec

Date

Fig. 10. Model calculation of the ventilation rate of 12 fattening pig rooms with a central ventilation system for a meteorologicalreference year (hourly averages); every 9 days a new batch is started in one of the rooms

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

Date

Cum

ulat

ive

amm

onia

em

issi

on p

er b

atch

, kg

(fat

teni

ng p

ig)−1

1 Jan 1 Feb 4 Mar 5 May 5 Jun 6 Jul 6 Aug 6 Sep 7 Oct 7 Nov 8 Dec4 Apr

Fig. 11. Model calculation of the ammonia emission from a fattening pig room, accumulated per batch

R.W. MELSE ET AL.78

the exhaust air from the central ventilation system and isshown in Fig. 12.

3.2.3. Effect of the use of a scrubber with a bypass vent

on the year-round ammonia emission of a pig house

Finally, from the modelled year-round ventilationrate and ammonia concentration data for the centralventilation system, the effect of the installation of an air

scrubber with a bypass vent was determined (Fig. 13), inthe same way as was done for the datasets with theactual ammonium and ventilation rate measurements(Figs 7 and 8). In Fig. 13, the ammonia load in kg h�1

that would be bypassed if an air scrubber with a bypassvent had been installed is plotted as a function of thebypass setpoint flow rate. As long as the tangent lineslopes of the curves in Fig. 13 decrease (get more

ARTICLE IN PRESS

Am

mon

ia c

once

ntra

tion,

mg

m−3

30

25

20

15

10

5

01 Jan 1 Feb 4 Mar 4 Apr 5 May 5 Jun 6 Jul 6 Aug 6 Sep 7 Oct 7 Nov 8 Dec

Date

Fig. 12. Model calculation of the ammonia concentration in the air from a central ventilation system for 12 rooms for ameteorological reference year (hourly averages); every 9 days a new batch is started in one of the rooms

Byp

asse

d am

mon

ia lo

ad, %

of

tota

l loa

d

Central ventilationsystem for 12 rooms

Ventilation per room

100

90

80

70

60

50

40

30

20

10

0 10 20 30 40 50 60 70 800

Bypass setpoint, m3 (fattening pig)−1 h−1

Fig. 13. Amount of ammonia that would be bypassed if a scrubber with a bypass vent was installed as a function of the bypasssetpoint, both for ventilation per room and for a central ventilation system; model calculations for a meteorological reference year

SIZE REDUCTION OF AMMONIA SCRUBBERS 79

negative) with a decreasing bypass setpoint flow rate, theefficiency of scrubber utilisation in kg [NH3 removal]m�3 [scrubber volume] increases, assuming that thesame relative ammonia removal (%) is achieved and thesame air loading rate of the scrubber in m3 m�3

[scrubber volume] h�1 is used as for a conventionalscrubber without a bypass vent.For a ventilation system per room, the simulation of

Fig. 13 shows that in case of a bypass setpoint flow rate

of 30m3 (fattening pig)�1 h�1, still 84% of the ammoniaload enters the scrubber and only 16% is venteduntreated. The datasets with the measured ammoniaemissions showed that in this situation 20% of theammonia was vented untreated (Fig. 7, average ofdataset 1–3), so the simulation of Fig. 13, which is ayear-round simulation using a meteorological referenceyear, indicates that the cost-reducing effect of using abypass vent is even stronger than was shown by Fig. 7.

ARTICLE IN PRESS

R.W. MELSE ET AL.80

This can be explained by the fact that each dataset withmeasurement was gathered during one batch only withthe specific weather conditions at that time and place,but could also by influenced by limitations of the modelthat was used.

Furthermore, Fig. 13 shows that the use of a centralventilation system, as had been expected, further reducesthe cost of ammonia removal by air scrubbing. In caseof a bypass setpoint flow rate of 30m3 (fattening pig)�1

h�1, still 89% of the ammonia load enters the scrubberand only 11% is vented untreated, whereas 16% wasvented untreated for the ventilation system per room.The efficiency of scrubber utilisation has its maximumvalue at a bypass setpoint of 13m3 (fattening pig)�1 h�1

or lower for the central ventilation system.Both the investment and operational costs of air

scrubbing will be reduced, as an increased scrubberutilisation efficiency means that a relatively smallscrubber needs to be operated in comparison with aconventional scrubber without a bypass venting system.For a bypass scrubber system, the investment cost perm3 scrubber size will be slightly higher than for aconventional scrubber system because of the additionalcosts of the bypass itself.

3.3. General discussion

In Figs 7, 8, and 13, it was demonstrated that for thetreatment of exhaust air from pig and poultry houses theuse of a bypass vent increases the efficiency of scrubberutilisation in kg [NH3 removal] m�3 [scrubber volume],as long as the maximum air loading rate of the scrubberin m3 m�3 [scrubber volume] h�1 remains unchangedand the scrubber size thus is reduced. This implies thatthe costs of air scrubbing will be reduced, as theinvestment and operational costs of air scrubbing arecorrelated to the scrubber size.

Bypassing peaks in the airflow results in an increase ofthe average air loading rate of the system, even when themaximum air loading rate of the scrubber is assumed to beunchanged. A conventional scrubber that has beendesigned for a maximum air flow of 60m3 (fatteningpig)�1 h�1 has an average air loading rate of1600m3m�3 [scrubber volume] h�1 and a maximum airloading rate of 3600m3m�3 [scrubber volume] h�1 (mini-mum EBRT is 1�0 s). From the data for the centralventilation system presented in Fig. 10, it can be calculatedthat the installation of a half-sized scrubber with a bypassvent at a bypass setpoint flow rate of 30m3 (fatteningpig)�1 h�1 increases the average air loading rate of thescrubber with 60% from 1600 to 2600m3m�3 [scrubbervolume] h�1, assuming the same maximum air loading rateis applied. This increase of the average air loading rate, or

air velocity in the scrubber bed, increases the pressure dropover the scrubber and thus increases the energy costs of themechanical ventilation system. A conventional scrubberwithout a bypass vent in general has a maximum pressuredrop of about 200Pa at the designed maximum airflowrate and an average pressure drop of about 50Pa; it wascalculated that the average pressure drop will increasefrom 50 to about 100Pa by the use of a smaller scrubberwith a bypass vent.

As it is desirable to minimise the extra pressure dropin the scrubber, the spatial dimensions of the scrubber(length, width, height) can be changed to level down thisincrease of the pressure drop, even if the scrubbervolume remains unchanged. When the surface area ofthe scrubber is increased by increasing the, in case of anupward airflow direction, width and length of thescrubber, the surface loading rate in m3 [air] m�2

[scrubber area] h�1 decreases and thus a lower airvelocity is achieved in the bed. At the same bed volume,increasing the width and length means that the height ofthe bed is reduced. The pressure drop, which isproportional to the bed height and the square of theair velocity, can thus be levelled down. The modelcalculations show that the height of the bed needs to bereduced by 27% in order to reduce the average pressuredrop of the scrubber with the bypass -system down tothe average pressure drop of a conventional scrubber,assuming the same maximum air loading rate is appliedfor both systems.

Finally, the average ammonia load of the scrubberwith a bypass vent, expressed as kg [NH3] m

�3 [scrubbervolume] h�1, is higher than for a conventional scrubberwithout a bypass vent. From the data for the centralventilation system in Fig. 13, it can be calculated that theinstallation of a scrubber with a bypass vent at a bypasssetpoint flow rate of 30m3 (fattening pig)�1 h�1

increases the average ammonia loading rate of thescrubber from 20 to 35 g m�3 [scrubber volume] h�1. Asa result of the increased ammonia loading rate, therelative ammonia removal in % of the scrubber systemmight decrease in theory.

For an acid scrubber, it is expected that the relativeremoval efficiency will hardly be influenced by anincrease of the average ammonia loading rate as thescrubber is operated at low pH so that the ammonia iscaptured by the acid very quickly and removed with thedischarge water.

For a biotrickling filter, however, the relative removalefficiency might be influenced by an increase of theaverage ammonia loading rate. After its transfer to theliquid phase, the ammonia must be biologically con-verted and therefore a larger amount of nitrifyingbiomass is needed in the scrubber system at a higheraverage ammonia loading.

ARTICLE IN PRESS

SIZE REDUCTION OF AMMONIA SCRUBBERS 81

On the other hand, currently operating conventionalammonia scrubbers for treatment of exhaust air fromanimal houses, both acid scrubbers and biotricklingfilters, have been designed to achieve sufficient ammoniaremoval under all operating conditions that may occurin practice, including incidentally high ammonia and airloading rates. Therefore these systems might be capableof successful ammonia removal at higher averageammonia and air loading rates without reduction ofthe relative ammonia removal.Further experimental research is necessary to demon-

strate if the design and cost calculations for air scrubber,based on the relationships of Figs 7, 8, and 13, need tobe corrected for these aspects. Also it is necessary togather practical experience in design, application andcontrol of bypass venting system. Finally, further workis required to determine if the air scrubbing strategydescribed in this paper is applicable for mechanicalventilated houses for other animal species than pigs andpoultry (e.g. veal calves) and how it influences odouremission from animal houses.

4. Conclusion

Conventional acid and biological air scrubbers thatare used for ammonia removal from exhaust air at pigand poultry houses have the capacity for treating theentire airflow at all times. As a result of the fluctuatingammonia emission pattern of this air, i.e. the time courseof airflow rate and ammonia concentration, for most ofthe time these air scrubbers are overdimensioned andunderloaded.By model calculations and analyses of measurements

datasets it was demonstrated that the application of an airscrubber that bypasses peaks of the airflow, will signifi-cantly decrease the required scrubber size while ammoniaemission levels are only slightly increased. In such asystem, most of the air is treated by the scrubber but a partof the air is vented untreated at times that the actualairflow exceeds a certain setpoint (e.g. in case the bypasssetpoint is set to 50% of the maximum airflow rate and thescrubber volume is reduced by 50%, only 10–20% of thetotal ammonia load is bypassed and subsequently emittedto the environment, for the analysed datasets and madeassumptions). This means that the use of the bypassventing system increases both the efficiency of scrubberutilisation in kg [NH3 removal] m�3 [scrubber volume] andthe cost-effectiveness of air scrubbing for ammoniaremoval in kg [NH3 removal] h�1.Although the use of a scrubber with a bypass venting

system decreases the maximum air in m3 m�3 h�1 andammonium loading rate in kg m�3 h�1, the average airand ammonium loading rate are increased in comparison

with a conventional scrubber. Further experimentalresearch on acid and biological scrubbers is necessaryto demonstrate to what extent this might affect theperformance of the scrubbers with regard to the relativeremoval efficiency in % and how this affects the totalemission from scrubbers with bypass systems.

Acknowledgements

This work was supported by the Dutch Ministry ofAgriculture, Nature and Food Quality (LNV). Wethank N.W.M. Ogink for his suggestions on themanuscript.

References

ASG (2004). Handboek pluimveehouderij. [Handbook ofpoultry farming.] Praktijkboek 36. Praktijkonderzoek Ani-mal Sciences Group, Lelystad, Netherlands, 244pp

Bruce J M; Clark J J (1979). Models of heat production andcritical temperature for growing pigs. Animal Production,28(3), 353–369

CBS (2004). StatLine. Statistics data bank for The Nether-lands. Available at: http://www.cbs.nl

Focht D D; Verstraete W (1977). Biochemical ecology ofnitrification and denitrification. Advances in MicrobialEcology, 1, 135–214

Groenestein C M; Huis in ‘t Veld J H W (1996). Praktijkon-derzoek naar de ammoniakemissie van stallen XXVII.Vleesvarkensstal met koeling van mestoppervlak in dekelder. [Research into the ammonia emission from livestockhousing systems XXVII: fattening pig house with slurrysurface cooling.] Report 96-1003. IMAG-DLO, Wagenin-gen, Netherlands

Heij G J; Erisman J W (1995). Acid rain research: do we haveenough answers? Proceedings of a Speciality Conference, ‘s-Hertogenbosch, The Netherlands, 10–12 October 1994.Studies in Environmental Science, 64, ISBN 0-444-82038-8

Heij G J; Erisman J W (1997). Acid atmospheric depositionand its effects on terrestrial ecosystems in The Netherlands.The Third and Final Phase (1991–1995). Studies inEnvironmental Science, 69, ISBN 0-444-82037-X

Hol J M G; Groenestein C M (2005). Unpublished data.Agrotechnology & Food Innovations, Wageningen UR,Wageningen, Netherlands

Hol J M G; Groot Koerkamp PWG (1998). Praktijkonderzoeknaar de ammoniakemissie van stallen XXXX. Vleeskuiken-stal met verwarming en koeling van de vloer met strooisel[Research into the ammonia emission from livestockhousing systems XXXX: broiler house with heating andcooling of litter covered floor.] Report 98-1004. IMAG-DLO, Wageningen, Netherlands

Huis in ‘t Veld J H W; Groenestein C M (1995). Praktijkon-derzoek naar de ammoniakemissie van stallen XXIV[Research into the ammonia emission from livestock housingsystems XXIV: fattening pig house with manure dilution bycollection in ammonia-free liquid.] Vleesvarkensstal metverdunning van mest door opvang in ammoniakvrije

ARTICLE IN PRESS

R.W. MELSE ET AL.82

vloeistof. Repport 95-1007. IMAG-DLO, Wageningen,Netherlands

Lund H (1984). Test reference years (TRY). Weather data setsfor computer simulations of solar energy systems and energyconsumptions. CEC Programme Solar Energy, ThermalInsulation Laboratory, Technical University of Denmark,Kongens Lyngby, Denmark

Melse R W; Mol G (2004). Odour and ammonia removal frompig house exhaust air using a biotrickling filter. WaterScience and Technology, 50(4), 275–282

Melse R W; Ogink N W M (2005). Air scrubbing techniquesfor ammonia and odor reduction at livestock operations:review of on-farm research in The Netherlands. Transac-tions of the ASAE, 48(6), 2303–2313

Melse R W; Willers H C (2005). Theoretical evaluation ofammonia scrubbers for governmental approval. In: Pro-ceedings of the 7th International Conference on Construc-tion, Technology and Environment in Livestock Farming,March 2–3, 2005, Braunschweig, Germany, pp 295–300,Kuratorium fur Technik und Bauwesen in der Land-wirtschaft e.V. (KTBL), Darmstadt, Germany, ISBN 3-7843-2185-2

Mosquera J; Hofschreuder P; Erisman J W; Mulder E; Van ‘tKlooster C E; Ogink N; Swierstra D; Verdoes N (2002).Meetmethoden gasvormige emissies uit de veehouderij.[Methods for measurement of gaseous emissions fromanimal husbandry.] Report 2002-12. IMAG-DLO, Wagen-ingen, Netherlands, ISBN 90-5406-216-9

Mosquera J; Hol J M G; Ogink N W M (2005). Analyseammoniakemissieniveaus van praktijkbedrijven in de var-kenshouderij (1990–2003). [Analysis of ammonia emissionlevels of pig farms (1990–2003).] Report 312. Agrotechnol-ogy & Food Innovations, Wageningen UR, Wageningen,Netherlands, ISBN 30-6754-874

Ni J-Q; Hendriks J; Vinckier C; Coenegrachts J (2000).Development and validation of a dynamic mathematicalmodel of ammonia release in pig house. EnvironmentInternational, 26, 105–115

Ogink N W M; Lens P N (2001). Geuremissies uit deveehouderij I. Overzichtsrapportage 1996–1999. [OdourEmissions from Animal Husbandry I. General report1996–1999.] Report 2001-14. IMAG-DLO, Wageningen,Netherlands

Prosser I J (ed.) (1986). Nitrification. Society for GeneralMicrobiology Special Publication 20. IRL Press, Oxford,UK, ISBN 1-85221-001-X and 1-85221-013-3

Satter I H G; Hol J M G; Huis in ‘t Veld J H W; Groenestein CM (1997). Praktijkonderzoek naar de ammoniakemissie vanstallen XXXIV. Vleesvarkensstal met mestverwijdering doorspoelen met dunne mestfractie via spoelgoten. [Researchinto the ammonia emission from livestock housing systemsXXXIV: fattening pig house with manure removal by aflushing-gutter system.] Report 97-1004. IMAG-DLO,Wageningen, Netherlands

Seedorf J; Hartung J; Schroder M; Linkert K H; Pedersen S;

Takai H; Johnsen J O; Metz J H M; Groot Koerkamp P W

G; Uenk G H; Phillips V R; Holden M R; Sneath R W; Short

J L; White R P; Wathes C W (1998). A survey of ventilationrates in livestock buildings in Northern Europe. Journal ofAgricultural Engineering Research, 70, 39–47

Sterrenburg P; Van Ouwerkerk E N J (1986a). Rekenmodelvoor de bepaling van de thermische behaaglijkheidszone vanvarkens (BEZOVA). [Calculation model for determinationof the thermal comfort zone of pigs (BEZOVA).] Report 78,IMAG-DLO, Wageningen, Netherlands

Sterrenburg P; Van Ouwerkerk E N J (1986b). Tabellen vanwarmteproduktie en temperatuureisen van varkens. [Tableswith heat production and temperature requirements of pigs.]Report 79. IMAG-DLO, Wageningen, Netherlands

Van Ouwerkerk E N J (1999). ANIPRO klimaat- enenergiesimulatiesoftware voor stallen. [ANIPRO climateand energy simulation software for animal houses.] ReportV99-109. IMAG-DLO, Wageningen, Netherlands

Van Wagenberg A V; Vermeij I (2001a). Energiegebruik enkosten van centrale afzuiging en afzuiging per afdeling invarkensstallen. [Energy use and costs of ventilation usingeither central fans or one fan per pig room.] Report 218.Praktijkonderzoek Veehouderij, Lelystad, Wageningen UR,Netherlands, ISSN 0169-3689

Van Wagenberg A V; Vermeij I (2001b). The use of a centralextraction ventilation system for pig facilities, energy savingsand economic consequences. ASAE Annual InternationalMeeting, Sacramento, CA, USA, July 30–August 1, 2001.ASAE Paper No. 01-4050

Wever A C; Huis in ‘t Veld J W H ; Groot Koerkamp P W G

(1999). Onderzoek naar de ammoniakemissie vanstallen XLV. Vleeskuikenstal met isolatie en ventilatievolgens het VEA-concept. [Research into the ammoniaemission from livestock housing systems XLV: broilerhouse with insulation and ventilation according to theVEA concept.] Report 99-09, IMAG-DLO, Wageningen,Netherlands