Burst Releases of Hydrogen Sulfide in Mechanically Ventilated Swine Buildings

Copyright (c) 2000 Water Environment Federation. All Rights Reserved.

Odors and VOC Emissions 2000

1

BURST RELEASES OF HYDROGEN SULFIDE IN MECHANICALLY VENTILATEDSWINE BUILDINGS

Ji-Qin Ni, Albert J. Heber, Claude A. Diehl, Teng T. LimR. K. Duggirala and B. L. Haymore

Agricultural & Biological Engineering Dept. Purdue University, 1146 ABE Building, West Lafayette, IN 47907

ABSTRACT

Hydrogen sulfide (H2S) is an odorous gas produced from animal manure. It is toxic to humans andanimals at high concentrations. There is little understanding about its release behavior from storedliquid swine manure. This paper documents burst releases of H2S recorded with a state-of-the-artfield measurement system in two commercial swine buildings over a six-month period. The 1,000-head grow-finish buildings had 2.4 m deep manure collection and storage pits under fully slattedfloors. They were mechanically ventilated with pit chimney fans and end wall exhaust fans. Pitchimney ventilation rates were continuously measured with full-size fan-impeller anemometers.End wall fan ventilation rates were calculated from continuously recorded fan operation times anddifferential static pressures. Concentrations of H2S in the sampled air streams were measured withan H2S converter and an SO2 analyzer for each building. Sample air was continuously pumpedfrom three locations: pit fans, wall fans and pit headspaces. Hydrogen sulfide concentrations ateach location were acquired for 10 or 15 minutes during each 60 or 90 minute sampling period,respectively.

A burst release of H2S was defined as a sudden increase (100% or greater) of H2S release ratemeasured during any period compared to previous periods, under relatively constant ventilationrates and indoor room temperatures. A total of 83 burst releases were identified in the data of 219days with reliable measurements. Releases of ammonia (NH3), which were simultaneously andcontinually measured, were verified to confirm that the burst H2S releases were unique. Typicalburst H2S releases were presented graphically. Time distributions of the bursts were studied. Theburst H2S releases were not related to any factors known to affect NH3 releases from liquidmanure. Further research is needed to explain the causes of the burst releases documented in thispaper.

KEYWORDS

Air quality, air pollution, agricultural emission, environment, ammonia, swine manure

INTRODUCTION

Hydrogen sulfide (H2S) is one of the main pollution gases emitted from animal facilities. It iscolorless but has a very pungent odor similar to rotten eggs. Concentrations of H2S are usuallyvery low in swine buildings compared with other gases like carbon dioxide (CO2) and ammonia(NH3). Typical reported concentrations ranged from 18 to 1100 ppb (Muehling, 1970; Heber et al.,1997; Ni et al., 1999a).



2

High H2S concentration is toxic to humans and animals. A 50 ppm concentration of H2S can causedizziness, irritation of the respiratory tract, nausea, and headache. It is considered immediatelydangerous to life and health at 300 ppm. Death from respiratory paralysis can occur with little orno warning at concentrations exceeding 1,000 ppm (Field, 1980). It has been responsible for manyanimal as well as human deaths in animal facilities (Field, 1980; Osbern and Crapo, 1981; Hagleyand South, 1983; Anonymous, 1996). Critically high concentrations of H2S in confined swinefacilities were usually related to agitation of manure during pump-out when large quantities of H2Swere released (Patni and Clarke, 1991) or when insufficient ventilation existed.

Although many reports of H2S related accidents and odor nuisance have attracted public attention,scientific research on H2S characteristics in animal facilities is far from satisfactory. Most of thestudies of H2S focused on concentrations measured with a wide variety of instruments andprocedures. The H2S concentration inside a confinement swine building is primarily influenced bythe H2S input from the source (stored manure) and the clearance of H2S to the outdoors byventilation. Scientific study of the H2S production by manure decomposition, the H2S release fromthe manure surface into free air, and the H2S emission from confined buildings to the atmospheretherefore deserve more attention. However, very little information about H2S release in swinehouses exists in the literature. Better understanding of the release behavior of H2S will benefitprofessional working to improve environmental protection, public health and farm safety.

Ni et al. (2000) reported two over-night experiments designed to study quantity and behavior ofgas releases under controlled ventilation and temperature in two mechanically-ventilated swinebuildings that had been emptied between pig groups. A strange behavior of “H2S burst release”was discovered. It was characterized by sudden increases of H2S release from manure in spite ofconstant building temperature and ventilation rates. Although the cause of this burst release is stillunknown, the following question remains to be answered: Does this phenomenon also occur inoccupied swine buildings?

OBJECTIVES

The objectives of this paper are to:

1) identify burst releases of H2S in two fully occupied swine finish buildings,2) compare burst releases of H2S with corresponding releases of NH3, and3) evaluate seasonal and diurnal distributions of H2S burst releases.

MATERIALS AND METHODS

Test Buildings and Measurement Systems

Data used in this paper were collected from two mechanically-ventilated, 1,000-pig finishingbuildings in Illinois, which were involved in a field test of a commercial manure additive (Heber etal., 1998a). The two buildings, designated “3B” and “4B,” were control buildings and thereforewere not treated with the additive. Building 3B was one of the two buildings, in which Ni et al.(2000) discovered the burst H2S release during empty-room experiments.



3

Each of the 12.3 m by 65.9 m buildings had a 2.4 m deep pit under a fully slatted floor. The pitsurface area was 799 m2 (Figure 1). There were four pit ventilation fans and five wall ventilationfans in each building. The four, 0.46m diameter, variable-speed pit fans, each installed at the topof a vertical chimney, operated continuously. The building was tunnel-ventilated during hotweather. One, 90-cm diameter and four, 120-cm diameter exhaust fans were located on the eastend wall of the building. Heber et al. (1998b) provided detailed descriptions of the buildings andtest instrumentation.

Figure 1 - Floor plan (top) and side view (bottom) of the buildings with measurement andsampling locations.

129'

96, 13

8

36' 18'

Wall fans

0.8

5.9

5.9

↑↑↑↑North 7.27.5

24.3 23.10.8

*

1.6Cleanout port

1.91.7

Pit fan gas sampling*Temperature

Pit headspace gas sampling

*2.32.4

2.4

65.9

Pressure measurementWall fan gas sampling

Attic

Manure pit

Pit fan chimneys

Unit: meter

Room

Air quality measurements in the two buildings started on March 6 and ended on September 25,1997. Hydrogen sulfide was measured with an SO2 analyzer (Model 45, Thermal EnvironmentalInstruments (TEI), Mansfield, MA) after being converted to sulfur dioxide (SO2) with an H2Sconverter (TEI Model 340). Ammonia was measured with an NH3 converter and an NH3 analyzer(TEI Model 17C). Each building used one set of instruments for H2S and NH3 measurement. Airwas sampled at three locations: 1) pit headspace (six sampling points), 2) pit fans (four samplingpoints), and 3) end wall fans (five sampling points). Multiple sampling points of each locationwere connected in parallel to the gas sampling system (Figure 1).

Only locations 1 and 2 were connected at the beginning of the test. Gas at each location wassampled and measured continuously for 15 minutes before switching to another location. Blocks ofdata collected at 1 second were averaged, and the average of 20 readings stored every 20 seconds.Thirty minutes were allocated during each 60-minute cycle to measure gas concentrations in eachbuilding. Thus, 24 sampling periods were obtained daily for gas concentrations at each location.



4

Gas location 3 was added in 3B on June 4 and in 4B on July 16. A 90-minute period was used witha sampling duration of 15 minutes at each location. Thus, 16 sampling periods were obtaineddaily. The sampling duration at each location was reduced from 15 to 10 minutes on August 14.Thirty minutes every hour were allocated for each building and the number of sampling periodsreturned to 24 each day. The H2S concentration data collected during the first three minutes ofeach 15 or 10-minute sampling duration were ignored to allow analyzers to equilibrate. Thus, 12or 7 minutes of useful gas concentration data were obtained during each period.

Building ventilation rate was the sum of the ventilation rates of the wall fans and of the pit fans.Ventilation rates from the wall fans were calculated based on fan curves supplied by themanufacturer and the building static pressure. A fan-impeller anemometer (FanCom Model FMS50, Techmark, Lansing, MI) was installed below the ventilation fan inside each pit chimney of 4B.The ventilation rate of the pit fans was the sum of the measured ventilation rates by the foursensors. There was no fan-impeller anemometer installed in 3B. The ventilation rate from the pitfans was estimated based on fan control voltage. The four chimney fans were controlled in parallelby a single speed controller. An airflow/voltage relationship was determined from other buildingsthat were equipped with pit fans and ventilation rate sensors. The pit fans operated at full speedduring most of this test.

Indoor air temperatures were measured at seven locations equally spaced along the center of eachbuilding length two meters above the floor with semiconductor sensors (Model AD592, ComputerBoards, Mansfield, MA). Outdoor air temperature was measured at a height of three meters with atemperature probe (Model EHRH/T1-2-I-1, General Eastern, Woburn, MA).

Pigs were weighed before entering the building. The average weight of the pigs before the firstgroup reached market weight was calculated based on beginning and ending pig weights andnumber of days in the building. Daily pig weight was calculated according to an empiricalrelationship for pigs with an average growth rate of 0.75 kg/d (van Ouwerkerk and Aarnink, 1992).This test covered parts of two pig growth cycles.

Identification of Burst Releases of Hydrogen Sulfide

There were days with incomplete air quality data, because of data acquisition problems and alightning strike. To reduce errors in the investigation, only days with complete and reliablemeasurement, including 124 days in 3B (from March 18 - September 24) and 95 days in 4B (fromApril 5 to September 25), were selected. Each day of complete measurement covered 24 hours andwas divided into 24 or 16 periods. Each period had one data subset that consisted of H2Sconcentrations at each location, ventilation rate, indoor and outdoor temperatures, pig number andpig weight.

The H2S emission rates were obtained by multiplying simultaneously measured ventilation rateswith H2S concentrations at the corresponding ventilation fans. Concentrations at pit headspaceswere used when H2S concentrations at the wall fans were not yet measured (before June 4 in 3Band before July 16 in 4B).. The gas release rate is different from the gas emission rate under non-steady state conditions. However, as an engineering approach, the H2S release rate wasapproximated by its emission rate in this study because it was technically difficult to directlymeasure the gas release rate in the manure pits.



5

The burst release of H2S was defined as a sudden increase (100% or greater) of H2S release ratemeasured during any period compared to previous periods, under relatively constant ventilationrates and indoor room temperatures.

The following steps were used to identify the H2S burst release:

1) Raw data from field tests were processed to obtain period means of measured variables. Dailyfiles with either 24 or 16 periods of data were generated.

2) Twelve graphs including gas concentrations, gas emissions, ventilation rates, and temperatureswere printed for each day.

3) The graphs were visually examined to find periods with sudden increase of H2S release rates ascompared with previous periods under relatively constant ventilation rates and indoortemperatures.

4) Data from visually selected burst H2S releases were calculated to confirm that the bursts wereat least 100% higher than the previous periods and that there were no significant changes inventilation rate and temperature during these periods.

5) Corresponding releases of ammonia (NH3), which were simultaneously and continuallymeasured, were inspected to confirm that the burst H2S releases were unique.

RESULTS

A total of 83 burst releases of H2S were identified. Forty-two burst releases were found during 37of 124 available days of data in 3B. There were five days with two bursts each day. Burst durationranged from 1 to 7 hours (Figure 2).

Figure 2 - Data availability, burst distribution and burst duration in 3B.

0

2

4

6

8

Bur

st du

ratio

n (h

)

0

3/18 4/8

4/29

5/20

6/10 7/1

7/22

8/12 9/2

9/23

Date in 1997

Ava

ilabl

e da

te



6

Forty-one burst releases were found during 38 of 95 days in 4B (Figure 3). There were one daywith two bursts and another day with three bursts. The duration of individual bursts in 4B rangedfrom 1 to 14 hours. Many were longer than those in 3B.

Hourly distribution of detected bursts showed two peaks, one between 03:00 and 07:00 andanother between 10:00 and 23:00 (Figure 4). The two peaks in 3B were more clearly defined.

Figure 3 - Data availability, burst distribution and burst duration in 4B.

0

5

10

15

Burs

t dur

atio

n (h

)

0

3/18 4/8

4/29

5/20

6/10 7/1

7/22

8/12 9/2

9/23

Date in 1997

Ava

ilabl

e da

te

Figure 4 - Hourly distribution of the burst releases in 3B and 4B.

0

4

8

12

16

20

0:00 3:00 6:00 9:00 12:00 15:00 18:00 21:00Time of day

Num

ber o

f bur

sts 3B 4B



7

A typical burst H2S release in 3B on September 13, 1997 is presented in Figure 5. There were 874pigs with an average weight of 72.5 kg each in the building. Release and concentration patterns ofNH3 under the same environmental conditions are also presented for comparison.

Figure 5 – Data collected in 3B on September 13, 1997 including hydrogen sulfide (topgraph), ammonia (middle graph), and ventilation rate and room temperature (bottomgraph). C-WF is concentration at wall fans; C-CH is concentration at pit chimneys; C-HS isconcentration in pit headspace; RE is release rate; RT is room temperature; OT is outdoortemperature; and QT is ventilation rate.

0

150

300

450

600

750

900

H2S

con

cent

ratio

n (p

pb)

0

30

60

90

120

150

180

H2S

rele

ase

(g/h

)C-WFC-CHC-HSRE

0

5

10

15

20

NH

3 con

cent

ratio

n (p

pm)

0

500

1000

1500

2000

NH

3 rel

ease

(g/h

)C-WF C-CHC-HS RE

0

5

10

15

20

25

30

0:00 3:00 6:00 9:00 12:00 15:00 18:00 21:00Hour of day

Tem

pera

ture

(°C

)

0

50

100

150

200

250

Ven

tilat

ion

(x10

00 m

3 )

RTOTQV



8

DISCUSSION

Production of H2S by decomposing liquid manure is related to microbiological and chemicalprocesses. It is affected by the quantity and quality of manure, fresh manure drop-in, temperatureand other factors that are ever changing in swine buildings. Dissolved H2S is released from theliquid manure and enters into the free air stream in the building. Release of H2S, like the release ofother gases, is believed to be a process of mass transfer across the liquid-gas interface. Hydrogensulfide emission is a process of H2S emanating from inside the building into the outdooratmosphere. Concentration of H2S inside a building is a state variable. It is governed by the H2Sinput (release from manure), output (emission to outdoors) and the air volume of the building.Moreover, H2S concentration gradients exist in large swine buildings (Ni et al., 1999a).

Hydrogen Sulfide Release and Hydrogen Sulfide Emission

It is technically difficult to measure the H2S release from manure in large manure storage pits.Measurements were usually conducted at the ventilation outlet to obtain H2S emission rates(Avery et al., 1975; Heber et al., 1997; Ni et al., 1999b). Two errors may have been introducedwhen H2S emissions were used to approximate H2S releases. One was the delayed response tochanges in H2S release rates due to transportation of H2S gas from the surface of the manure to themeasurement location. Each test building had an air volume of about 2900 m3 when the manure pitwas half full (1.2 m depth). It required only 3.5 and 0.7 minutes, compared to the air samplingperiods of 60 or 90 minutes, to ventilate 2900 m3 of air when the ventilation rates were 50,000 and236,000 m3/h, respectively (Figure 5). The pit chimney and pit headspace measurement locationswere both inside the pit and close to the source manure. Therefore, the potential delay would notappear to significantly affect identification of burst releases.

Another error may have been due to the room air volume acting analogous to a capacitance in anelectrical circuit. Some higher frequency and shorter duration burst releases may have beenconcealed because of this capacitance. The undetectable higher frequency burst releases may alsohave been concealed by the limitation of the 60 or 90 minutes air sampling periods and the 7 or 12minutes data average time.

Temporal Distribution of the Bursts

Burst releases of H2S were identified during relatively warm weather in this study. Releasessimilar to bursts during cold weather were also reported by Ni et al. (1998). Thus, the phenomenonmight not be restricted to certain seasons of the year. However, the daily distribution of identifiedbursts (Figures 2 and 3) did not exhibit an accurate picture of the bursts. This is because, in thefield, the bursts were only visible during relatively constant environmental conditions andventilation rates. Fluctuating ambient temperature affected indoor temperature and triggered theautomatic adjustment of ventilation rate in the buildings by electronic environmental controllers.Therefore, burst H2S releases were not identifiable during days and hours with unstable ambienttemperatures. Furthermore, data were sometimes not available due to equipment down time, e.g.between June and July for 4B (Figure 3).

The hourly burst distribution in Figure 4 demonstrated that bursts happened at all hours of the day.Similarly, Figure 4 is not intended to give a complete picture of time distribution. The two time



9

peaks in the figure corresponded to times when the ventilation rates were relatively constant andbursts could be relatively easily identified.

Location of Burst Releases in the Manure Pits

There was no evidence that sudden releases of H2S gas from manure occurred simultaneously andequivalently over the entire pit surface areas of 799 m2. It seems plausible that burst releasesoccurred at different locations and at different times. It also seems possible that unevenlydistributed temporal and spatial releases from the stored manure sometimes formed release peaksthat were identified as bursts in this study. A closer investigation of the particular phenomenon ofH2S release will require controlled laboratory tests with small manure surface areas andcontinuous monitoring of H2S release.

The concentration differences between the pit chimneys and the pit headspace, both sampled underthe floor in the pit, appear to exhibit the uneven release in the manure pit (Figure 5). Theconcentration differences of H2S as well as NH3 at different sampling locations were less visibleafter 09:00 as shown in Figure 5. This was probably due to higher ventilation rates during the day,which provided better air mixing.

Causes and Consequences of Burst Releases

Hydrogen sulfide and NH3 were both released from the manure in the pits. The main factorsaffecting the release process of NH3 are known to be the NH3 concentration difference betweenmanure and free air stream above the manure, temperature and airflow velocity over the manuresurface (Ni, 1999).

A comparison of the releases of H2S and NH3 revealed different behavior characteristics betweenthe two gases (Figure 5). Higher ventilation rates starting from 09:00 increased the release of bothgases. This could be explained by increased air velocity over the manure surface, which increasedthe mass transfer rate across the liquid-gas interface. The slight increase of room temperaturecould also accelerate the rate of mass transfer. However, a sudden increase of H2S releaseappeared at 17:00 when the ventilation rate, room temperature and release of NH3 remained ratherconstant. This demonstrated that the burst H2S release was unique. The airflow rate andtemperature did not seem to directly cause the burst H2S releases. Increased gas concentration inbuilding air would theoretically retard gas release from manure, because it reduces the gasconcentration difference between the manure surface and the room air. Would it be possible thatthere was a sudden increase of H2S concentration in the surface manure during the burst release, orthere were some yet unknown factors that triggered a chain reaction and suddenly accelerated therelease process of H2S? These questions require further investigation before they can be answered.

From the viewpoint of air quality, the importance of burst releases from manure was that theycaused sudden increases of H2S concentrations in swine buildings. In cases of insufficientventilation or ventilation failure, these bursts may worsen the air quality in the swine facilities andcreate a greater health hazard than indicated by long-term integrated averages. Moreover, burstreleases may introduce large measurement errors when average H2S data (e.g. daily average) are



10

required, but only spot and shorter duration (e.g. several minutes) measurements are conducted inswine facilities.

CONCLUSIONS

The following conclusions were drawn from this study:

1. Burst releases of H2S not only existed in emptied swine buildings as it was reported inprevious work, but also existed when the buildings were occupied with pigs.

2. The duration of a single burst ranged from 1 to 14 hours with the 60 or 90 minutes of gassampling periods. More bursts with shorter durations may have existed but could not beidentified under field conditions.

3. The burst releases of H2S were identified from March through September, but there was noevidence that they were limited to warm weather. Burst releases occurred at all hours of theday.

4. The burst H2S releases could not be related to ventilation or temperature changes, whichtypically affect NH3 releases from liquid manure. The particular phenomenon of H2S releaseneeds further laboratory investigation under better-controlled environments and restrictedmanure surface area.

5. The burst H2S releases caused increased H2S concentrations in the swine buildings. In case ofinsufficient ventilation or ventilation failure, they may worsen the air quality in the swinefacilities and therefore contribute to a potential health hazard. They may also introduce largeerrors when short duration and spot H2S measurements are performed in swine facilities.

ACKNOWLEDGEMENTS

This research was supported, in part, by the Multi-State Consortium on Animal Waste, the PurdueUniversity Agricultural Research Program, and Monsanto Enviro-Chem in St. Louis, MO. Theauthors also acknowledge the collaboration and assistance of Mr. Brad Begolka, Heartland Pork,Inc.

REFERENCES

Anonymous (1996). Nightmare on the hog farm: Hydrogen sulfide claims two lives in Minnesota.The Biobulletin. Vol. 3 No. 2. pp.1-3.

Avery, G. L., Merva, G. E. and Gerrish, J. B. (1975). Hydrogen sulfide production in swineconfinement units. Transactions of the ASAE. Vol. 18. pp.149-151.

Field, B. (1980). Rural health and safety guild: Beware of on-farm manure storage hazards.Cooperative Extension Service, Purdue University, West Lafayette, Indiana. S-82. 3 p.

Hagley, S. R. and South, D. L. (1983). Fatal inhalation of liquid manure gas. Medical Journal ofAustralia. Vol. 2 No. 9. pp.459-460.



11

Heber, A. J., Duggirala, R. K., Ni, J. Q., Spence, M. L., Haymore, B. L., Adamchuk, V. I., Bundy,D. S., Sutton, A. L., Kelly, D. T. and Keener, K. M. (1997). Manure treatment to reduce gasemissions from large swine houses. In: International Symposium on Ammonia and Odour Controlfrom Animal Production Facilities. Vinkeloord, The Netherlands, Oct. 6-10. Voermans, J. A. M.and Monteny, G. J. (Eds). Rosmalen, The Netherlands. Vol. II. pp.449-458.

Heber, A. J., Ni, J. Q., Lim, T. T. and Sutton, A. L. (1998a). Field tests of Alliance. Final report toMonsanto EnviroChem. Purdue University, West Lafayette. September. 266 p.

Heber, A. J., Ni, J. Q., Haymore, B. L., Duggirala, R. K., Diehl, C. A. and Spence, M. L. (1998b).Measurements of gas emissions from commercial swine buildings. In: ASAE Annual InternationalMeeting. Orlando, Florida, July 11-16. Paper No. 984058. ASAE, St. Joseph, MI. 40p.

Muehling, A. J. (1970). Gases and odors from stored swine waste. Journal of Animal Science. Vol.30. pp.526-530.

Ni, J. Q. (1999). Mechanistic models of ammonia release from liquid manure, a review. Journal ofAgricultural Engineering Research. Vol. 72 No. 1. pp.1-17.

Ni, J. Q., Lim, T., Heber, A. J., Diehl, C. and Duggirala, R. (1998). Gas release from deep manurepits in cleaned and emptied swine finishing buildings. In: Animal Production Systems and theEnvironment. An International Conference on Odor, Water Quality, Nutrient Management andSocioeconomic Issues. Des Moines, Iowa, July 20-22. Iowa State University, Des Moines, Iowa.Vol. I. pp.321-326.

Ni, J. Q., Heber, A. J., Lim, T. T. and Diehl, C. A. (1999a). Characteristics of hydrogen sulfideconcentrations in mechanically ventilated swine buildings. In: ASAE/CSAE-SCGR AnnualInternational Meeting. Toronto, July 18-22. Paper No. 994155. ASAE, St. Joseph, MI. 25p.

Ni, J. Q., Heber, A. J., Lim, T. T. and Diehl, C. A. (1999b). Continuous measurement of hydrogensulfide emission from two large swine finishing buildings. In: ASAE/CSAE-SCGR AnnualInternational Meeting. Toronto, July 18-22. Paper No. 994132. ASAE, St. Joseph, MI. 17p.

Ni, J. Q., Heber, A. J., Diehl, C. A. and Lim, T. T. (2000). Ammonia, hydrogen sulphide andcarbon dioxide from pig manure in under-floor deep pits. Journal of Agricultural EngineeringResearch, In press.

Osbern, L. N. and Crapo, R. O. (1981). Dung lung: a report of toxic exposure to liquid manure.Annals of Internal Medicine. Vol. 95 No. 3. pp.312-314.

Patni, N. K. and Clarke, S. P. (1991). Transient hazardous conditions in animal buildings due tomanure gas released during slurry mixing. Applied Engineering in Agriculture. Vol. 7 No. 4.pp.478-484.

van Ouwerkerk, E. N. J. and Aarnink, A. J. A. (1992). Gas production of fattening pigs. In:International Conference on Agricultural Engineering. Uppsala-Sweden, June 1-4. Paper No. 920207. 16p.

Date post:	28-Nov-2023
Category:	Documents
Upload:	independent
View:	1 times
Download:	0 times

Burst Releases of Hydrogen Sulfide in Mechanically Ventilated Swine Buildings

Documents