Well insulated walls of residences experience temperature depression in their outer layers during cold weather, causing moisture to condense on the surfaces. A predictive model capable of identifying the conditions that potentially lead to condensation or high moisture levels has been developed. The model utilized includes both moisture storage and distribution effects by utilizing the general form of the thermal energy and moisture conservation equations for each layer of a wall of typical residential construction, utilizing materials such as wood siding, insulation, and gypsum board.

Over the past decade many models have been developped to analyse thermal behaviour of buildings, but thermal comfort of occupants, which should never be forgotten, is seldom treated. The aim of this paper is to present three models created to get a complete tool to simulate thermal behaviour of a man in a real environment. The simplest one is the "PMV-PPD", which calculates comfort indices.

The "object oriented programming" and model reduction tniques give some new possibilities to develop computer tools. It is now possible to design a building on the computer, using computer objects corresponding to architectural concepts (materials, walls, windows .... ). Representing the building as a structure of objects is an approach which is particularly adapted to a thermal analysis and a comparison of designs, possibly with the help of an expert interface.

In recent years, there has been much discussion about the need for improvement in the quality of building design software. One area of design software improvement which has received much attention is that of information integration. Such integration is recognised as both desirable in its own right and necessary for advances in other areas of building design research. Advanced functions such as project management, intelligent design interfaces and complex building thermal performance simulation all depend to some extent on the free interchange of information.

The rapid development in the thermal energy modelling requirements for buildings, marked by the need to integrate many phenomena, has led the Applications de l'Electricit department at Electricit de France to develop a general energy simulation tool called CLIM 2000. Beyond the production of the software, our approach is to provide the specialists in the 7arious fields involved in the creation of the model library, with common formalisation rules ensuring clear and unambiguous expression of their work.To do this, we have drawn up a method based on a thermodynamic approach to the phenomena.

This paper describes two simulation software packages which permit building designers to understand how buildings will perform: the ESP building energy simulation system and the ARIA Computational Fluid Dynamics (CFD) air distribution simulation system. One of the major problems with CFD code is the specification of boundary conditions for the problem. ESP can provide the boundary conditions information for the CFD airflow, simulation. Two brief case studies are presented which illustrate the ability to provide the boundary conditions for the CFD problem from ESP.

The present means of building representation, fall short in establishing a common modelling base for the various application specific analysis and simulation programs, describing building behavior and performance. A "generic", object-oriented, approach to product modelling allows multiple design representations to be described as different views of a common, gradually evolving, building product model. The productmodel provides the capability to generate, in successive design iterations, a coherent description of the form, structure and dimensions of the building.

The aim of our paper is to present a multi-discipline CAD system named CONCEPTOR which allows the user to work during the various stages of the building engineering design. Most of the actions are dedicated to the building performances assessment for various technical domains with an extension of tools towards: economical estimating, quality analysis, checking of the solutions according to the regulations, the standards and the constraints established by the designer, coherence of the various works and solutions.

A simplified procedure is described, to model daylighting, cooling, and heating impacts of vertical glazing in commercial buildings. Both annual and peak impacts are calculated, as well as cooling-equipment -sizing impacts. Simple economic analyses (present worth and simple payback) are included. The name BEEM is used, for Building Energy Estimation Module. The public domain procedure is implemented in spread-sheet software, available at nominal charge for materials and handling.

In this paper, we want to show an application of fuzzy control to building thermal regulation. Thanks to a new learning method, inspired from connexionnist techniques, the controller learns to identify the rules it must use, either without any previous knowledge, or with approximative rules. This method is tested in simulation on a thermal regulation problem and is compared to a discrete time PI controller.