INTRODUCTION

Atria involve the following two natural phenomena:

It has been seen that the behavior of an atrium in cold climates is also mainly due to its ability to retain solar heat within it with the green house effect¹ and hence the existing research is also directed more towards 'solar distribution' within the atrium. This is typically distributed based on a formula taken from empirical data for different forms and conditions. With the help of computer programs, the total amount of energy coming into an atrium can be calculated, this can then be distributed to the various surfaces and 'light elements' (air), and some fairly close thermal predictions can be made². Still there is a lack of clarity in predicting the following in atria in anything but the standard box:

• Stratification of temperature
• Natural ventilation effects
• Comfort (Radiation, air flow patterns, temperature stratification andhumidity)

WARM CLIMATE CALCULATIONS

The above observations indicate that there is a need to study atrium temperature in warmer climates, especially in the area of temperature stratification and air movement within. There is also a need to find out ways of predicting behavior of atriums with regards to the above issues, using computer programs or

otherwise. This means that one has to find not only a mathematical or empirical model to predict temperature stratification and air movement, but also compare the results with real observations to make sure such predictions are reliable.

This paper aims to study temperature stratification and air movement in an atrium by comparing measured data and a computational fluid dynamics (CFD) model for the same atrium.

METHODOLOGY

• An appropriate building is selected to be the 'model atrium' for data collection and observation. It is important that the model atrium is close to the generic definition of an atrium, with respect to form, atrium-volume to building-volume ratio and in terms of its occupancy and usage. The Bradbury Building in Downtown Los Angeles satisfies most of the above conditions. See Figure 1.1.

• An appropriate method to collect temperature data is to be determined. The the efficiency of the data collectors will be determined by its size, ease of setup, ease of downloading information and ability to be synchronized in time. Keeping this in mind, data was collected with the help of Stowaway Temperaturedata loggers over a period of time with variations in weather and temperature conditions. See Figure 1.2.
• This data has to be organized in various forms in order to understand the behavior of the building itself. Whether stratification happens? If yes, then how and when? What is its magnitude? This data is then further filtered to provide for certain specific base-cases and/or worst cases to be used in the CFD model.

• A basic understanding of CFD procedures and selection of a simplistic input and graphic output oriented CFD program is undertaken. PHOENICS a CFD software developed by Concentration, Heat and Momentum Ltd. (CHAM) is used for the purposes of the CFD modeling. See Figure 1.3.
• The base-case is input and outputs are studied for its compliance to the real conditions. This is done by comparing the simulated stratification in the atrium with the measured one with the same boundary conditions.
• Consecutive runs for other conditions with different air inlet and outlet positions and mass flux variations of air into the atrium are carried out.

The primary conclusions will be based on the temperature data collected. One has to see whether the atrium behaves more or less as expected. What is the degree of stratification? Observations have to be made with respect to atrium's response over a period of time; daily or seasonally to the changing external conditions. The next set of conclusions would be based on the performance of the CFD model and its comparison to the 'real' situation. It will also be attempted to see whether the CFD model is appropriate to predict other conditions which cannot be necessarily verified with empirical data.

One of the most important primary observations based on just the data collected was that, the internal temperature fluctuations are clearly driven by the outside temperature variation, with no or very little time lag. The internal air temperature does not fall below a certain level even though the outside temperature decreases considerably. It is seen that with decreasing outside temperature there is a decrease in the magnitude of temperature differential. This can be a result of the buffer effect of the atrium. This is probably due to the greenhouse effect and lack of proper ventilation the atrium retains re-radiated heat from the adjoining spaces and does not allow the internal air temperatures to drop.The secondary conclusions which are drawn on the basis of the CFD model show that the measured conditions in the atrium are fairly close when simulations are carried out for an unventilated model, there is strong vertical stratification with very little or no internal air movement but simulations carried out with very small air influx into the space show considerable change in the air temperature distribution in the space, with a decrease and almost dissapearance of vertical stratification.

REFERENCES