Natural ventilation is an effective method to save energy required to condition buildings and to improve indoor air quality. Computational fluid dynamics (CFD) was used to model single-sided buoyancy-driven natural ventilation in a single room with a heater and door. The velocity and temperature profiles at the doorway agreed fairly well with published literature that includes Mahajan's experimental [2] and Schaelin et al's numerical studies [1]. The 2D and 3D models predicted the neutral level with a difference of 5.6 % and 0.08 % compared to the experimental results, respectively. Using solutions at the doorway, heat transfer rates were calculated. More realistic situations were studied considering conduction, various ambient conditions, wind speeds, and additional heat sources and furniture in the room. The heat loss through the wall was modeled and the airflow and temperature within the room showed no significant changes despite modeling conduction through the walls. Various ambient temperatures and wind speeds were tested, and the neutral level height and total heat transfer rate through the doorway increased with decreasing ambient temperatures. However, the neutral level did not significantly change as wind speeds varied. Total heat transfer rate at the doorway became positive, that is heat transferred into the room, with wind speed. Lastly, the effect of additional heat sources (mini-refrigerator, monitor and computer) and furniture (bookshelf, desk, chair and box) on airflow and heat transfer in the room was analyzed by comparing with a simple case of a room with a heater. Large velocities and high temperatures were predicted in the vicinity of the heat sources. However, the spatially averaged velocity and temperature did not change significantly despite additional heat sources. The room with furniture was modeled at lower ambient temperature, where the spatially averaged velocities were larger and temperatures were lower than the simple case. The room heated up and reached its thermal comfort level, but the velocities exceeded the maximum acceptable level set by ASHRAE guidelines [8]. Wind was considered simultaneously with the lower temperature, and the room was cooled faster with wind. However, the room was never able to achieve the comfortable level both in velocity and temperature. / Master of Science

Natural ventilation is the dominating ventilation process in ancient buildings like churches, and also in most domestic buildings in Sweden and in the rest of the world. These buildings are naturally ventilated via air infiltration and airing. Air infiltration is the airflow through adventitious leakages in the building envelope, while airing is the intentional air exchange through large openings like windows and doors. Airing can in turn be performed either as single-sided (one opening) or as cross flow ventilation (two or more openings located on different walls). The total air exchange affects heating energy and indoor air quality. In churches, deposition of airborne particles causes gradual soiling of indoor surfaces, including paintings and other pieces of art. Significant amounts of particles are emitted from visitors and from candles, incense, etc. Temporary airing is likely to reduce this problem, and it can also be used to adjust the indoor temperature. The present study investigates mechanisms and prediction models regarding air infiltration and open-door airing by means of field measurements, experiments in wind tunnel and computer modelling. In natural ventilation, both air infiltration and airing share the same driving forces, i.e. wind and buoyancy (indoor-outdoor temperature differences). Both forces turn out to be difficult to predict, especially wind induced flows and the combination of buoyancy and wind. In the first part of the present study, two of the most established models for predicting air infiltration rate in buildings were evaluated against measurements in three historical stone churches in Sweden. A correction factor of 0.8 is introduced to adjust one of the studied models (which yielded better predictions) for fitting the large single zones like churches. Based on field investigation and IR-thermography inspections, a detailed numerical model was developed for prediction of air infiltration, where input data included assessed level of the neutral pressure level (NPL). The model functionality was validated against measurements in one of the case studies, indicating reasonable prediction capability. It is suggested that this model is further developed by including a more systematic calibration system for more building types and with different weather conditions. Regarding airing, both single-sided and cross flow rates through the porches of various church buildings were measured with tracer gas method, as well as through direct measurements of the air velocity in a porch opening. Measurement results were compared with predictions attained from four previously developed models for single‑sided ventilation. Models that include terms for wind turbulence were found to yield somewhat better predictions. According to the performed measurements, the magnitude of one hour single-sided open-door airing in a church typically yields around 50% air exchange, indicating that this is a workable ventilation method, also for such large building volumes. A practical kind of diagram to facilitate estimation of suitable airing period is presented. The ability of the IDA Indoor Climate and Energy (IDA-ICE) computer program to predict airing rates was examined by comparing with field measurements in a church. The programs’ predictions of single-sided airflows through an open door of the church were of the same magnitude as the measured ones; however, the effect of wind direction was not well captured by the program, indicating a development potential. Finally, wind driven air flows through porch type openings of a church model were studied in a wind tunnel, where the airing rates were measured by tracer gas. At single-sided airing, a higher flow rate was observed at higher wind turbulence and when the opening was on the windward side of the building, in agreement with field measurements. Further, the airing rate was on the order of 15 times higher at cross flow than at single-sided airing. Realization of cross flow thus seems highly recommendable for enhanced airing. Calibration constants for a simple equation for wind driven flow through porches are presented. The measurements also indicate that advection through turbulence is a more important airing mechanism than pumping.   The present work adds knowledge particularly to the issues of air infiltration and airing through doors, in large single zones. The results can be applicable also to other kinds of large single-zone buildings, like industry halls, atriums and sports halls. / Naturlig ventilation är den dominerande ventilationsprocessen i äldre byggnader såsom kyrkor, och även i de flesta småhus i Sverige och övriga delar av världen. Luftinfiltration och vädring utgör viktiga komponenter i naturlig ventilation, där luftinfiltration är luftflöde genom oavsiktliga läckage i byggnadsskalet, medan vädring är avsiktligt luftutbyte genom stora öppningar såsom fönster och dörrar/portar. Vädring kan i sin tur ske ensidigt (genom en öppning) eller som tvärdrag (genom två eller flera öppningar belägna på olika ytterväggar). Det totala luftutbytet påverkar värmeförluster och inomhusluftens kvalité. I kyrkor orsakar avsättning av luftpartiklar en gradvis nedsmutsning av invändiga ytor, inklusive väggmålningar och andra konstföremål. Betydande mängder partiklar avges från besökare, tända ljus, rökelse, o.d. Tillfällig vädring kan minska detta problem, men även användas för att justera innetemperaturen. Föreliggande studie analyserar mekanismer och predikteringsmodeller gällande luftinfiltration och dörrvädring genom fältmätningar, vindtunnelförsök och datorsimuleringar. Luftinfiltration och vädring har samma drivkrafter, d.v.s. vind och termik (inne‑ute temperaturskillnader). Båda dessa drivkrafter är svåra att predicera, särskilt vindinducerade flöden och kombinationen av termik och vind. Två av de mest etablerade modellerna för luftinfiltrationsprediktering i byggnader har utvärderats via mätningar i tre kulturhistoriska stenkyrkor i Sverige. En korrigeringsfaktor av 0,8 föreslås för bättre prediktion av den ena modellen (som gav bäst resultat) gällande höga en-zonsbyggnader såsom kyrkor. En detaljerad numerisk modell är utvecklad för luftinfiltrationsprediktering, där indata baseras på fältundersökningar, inkl. IR-termografering och uppmätt av neutrala tryckplanet (NPL). Modellens funktionalitet har validerats via mätningar i en av fallstudierna och pekar på tämligen god prediktionsprestanda. Vidare utveckling av modellen föreslås, inkl. ett mer systematiskt kalibreringssystem, för olika typer av byggnader och väderförhållanden. Gällande vädring mättes både ensidigt flöde och tvärdrag genom portar i olika kyrkobyggnader med hjälp av spårgas samt direkta lufthastighetsmätningar i portöppning. Mätresultaten jämfördes med erhållna prediktioner från fyra tidigare utvecklade modeller för ensidig ventilation. De modeller som tog hänsyn till vindturbulens gav något bättre resultat. Enligt utförda mätningar medför en timmes ensidig portvädring i en kyrka cirka 50 % luftutbyte, vilket indikerar att detta är en tillämpbar ventilationsmetod, även för så pass stora byggnadsvolymer. Ett särskilt vädringsdiagram presenteras, som syftar till att underlätta uppskattning av erforderlig vädringsperiod. Vidare studerades predikteringsprestanda hos IDA Indoor Climate and Energy (IDA-ICE) simuleringsprogram avseende vädring, där simuleringsdata jämfördes med fältmätningar i en kyrka. Programmets prediktion av ensidigt luftflöde genom en öppen kyrkport var av samma storlekordning som det uppmäta; dock klarade programmet inte av att hantera inverkan av vindriktning så väl, vilket pekar på en utvecklingspotential. Avslutningsvis undersöktes vinddrivet flöde igenom portöppningar i en kyrkmodell i vindtunnel, där luftomsättningen mättes med hjälp av spårgasmetoden. Vid ensidig vädring observerades högre flöde vid högre vindturbulens och när öppningen var på vindsidan av byggnaden, i överensstämmelse med fältmätningarna. Dessutom var vädringsflödet vid tvärdrag i storleksordningen 15 högre än det vid ensidig vädring. Det verkar alltså som att man kan öka vädringstakten avsevärt om man kan åstadkomma tvärdrag. Kalibreringskonstanter presenteras också för en enkel ekvation för vinddrivet flöde genom portar. Vindtunnelstudien indikerar vidare att advektion genom turbulens är en viktigare vädringsmekanism än pumpning. Föreliggande arbete bidrar med kunskap speciellt kring luftinfiltration och vädring genom portar i höga en-zonsbyggnader. Resultaten kan även vara tillämpliga på andra typer av höga en-zonsbyggnader såsom industrihallar, atrier/ljusgårdar och idrottshallar. / Church project