The air flow in a doorway is governed by density difference caused by temperature difference and pressure difference caused by mechanical ventilation. Tests have been carried out in a unique indoor test house where the room to room to temperature difference could be controlled very accurately with a new control system. In addition to these tests some tests were carried out in a scale model with water as the operating fluid. Two main criteria of unidirectional flow in a doorway have been explored:

This paper deals with the convective flow through a horizontal aperture connecting two superimposed large enclosures which are kept at different temperatures. The lower room is warmer than the upper room and this unstable thermal configuration generates a natural them1osyphon flow, between both volumes. This type of flow can occur inside buildings: stairwell flows or natural ventilation flows through horizontal openings. In the literature, very little information is available concerning this domain of applications.

A problem in a livestock building ventilated in the system with wall inlet at two sides is wind effects to the indoor air flow patterns. The in-proper inlet jet penetration profiles may cause draught at occupied zone and unnecessary stress of animals. The wind effects on such a system can be reduced by applying windbreak to the inlets from design point of view. However, a remain issue is if the effect could be reduced by regulating inlet opening and how to perform the control operation.

This paper introduces a research programme investigating the application of CFD to large scale industrial premises. A number of modelling issues and two case studies are discussed. The research programme will lead to an increased degree of confidence of CFD simulation results in complicated environments.

A series of CFD and model experiments were carried out in order to find the most effective ventilation system in a separated refuse disposal facility. The ventilation system needed in the facility protects the working space from dust and odors generated by handling refuse. The desired ventilation system is to introduce the outdoor air from the one side of the working area and to exhausts the contaminated air through the opposite side of the refuse stock yard, so-called the unidirectional airflow ventilation.

The keeping of animals in livestock buildings requires the ventilation of these buildings. Good climate conditions for the animals have to be guaranteed as well as little emissions and immissions from the livestock buildings. An important role for both of these opposite requirements plays the airstream inside and outside of the building. For this reason wind tunnel modelling and digital imaging by using a laser light sheet is applied beside other methods at the Institute of Agricultural Engineering Bornim.

In modem livestock buildings the design of the ventilation systems is important in order to obtain good air distribution. The use of Computational Fluid Dynamics for predicting the air flow and air quality makes it possible to include the effect of room geometry, equipment and occupants in the design of ventilation systems. However, it is not appropriate to include the detailed geometry of a large group of lying or standing animals affecting the air flow in the building. It is necessary to have relatively simple models of the animals, which are easier to implement in the computer models.

Under the conditions that the air in each room is in the state of perfect mixing and ventilation is in steady state, a method to estimate steady state concentration distribution to match an arbitrary contaminant generation distribution is proposed, using concentration data obtained from short-time tracer gas experiment. Also, a method is proposed, which is used to estimate ventilation rate by adding other available equations.

A data acquisition system which uses a computer provides a more useful analysis system. Since the processing speed of computer is continuously increasing, the information than it is possible using conventional data acquisition systems. However, the raw measurements also include the signal noise which may lead to difficulty when the signal is analyzed. This work assesses an algorithm for removing possible signal noise, usually with high-frequency, from the measurement of tracer gas concentration.