English

In recent years, highly glazed atriums are favorable to architectural aesthetics and to taking advantage of daylighting and solar heating. However, t he estimation of the building load of an atrium building is difficult because of the complex thermal phenomena occurred in the atrium space. The study aims to find out the methods of conducting accurate simulations of the cooling loads of various types of atriums, using whole building energy simulation tool – EnergyPlus. Cases of atrium buildings are collected and classified into various categories.

Computational fluid dynamics (CFD) can provide detailed information of flow motion, temperature distributions and species dispersion in buildings. However, it may take hours or days, even weeks to simulate airflow in a building by using CFD on a single central processing unit (CPU) computer. Parallel computing on a multi-CPU supercomputer or computer cluster can reduce the computing time, but the cost for such high performance computing is prohibitive for many designers. Our paper introduces high performance parallel computing of the airflow simulations on a graphics processing unit (GPU).

Advanced natural ventilation (ANV), often characterised by the use of dedicated ventilation stacks, shafts and other architecture features such as atria, light wells, has gained popularity for natural ventilation design in recent decades. In this research, a prototype ANV system is proposed, and the likely thermal performance in a range of UK climatic conditions predicted using dynamic thermal simulation.

The aim of this paper is to present an automatic generator of zonal models that makes it possible to free the user from the choice of the specific flows models that have to be implemented in the zonal model and from the partitioning stage. The dynamic simulation tool called “O-Zone” is based on an advanced formulation of zonal models. It uses on a new way of sub-dividing the room that allows us to obtain a dynamic partitioning based on airflow patterns. The partitioning method chosen is the Octree method.

In the last years many building designers have turned their attention to natural ventilation, due to the potential benefits in terms of energy consumption related to ventilation and air-conditioning, especially in mild and moderate climates. Consequently, several calculation techniques have been developed to design and predict the performance of natural ventilation.  This article presents a review of the existing approaches to predict natural ventilation performance, including simple empirical models, nodal models (mono-zone and multi-zones), zonal models and CFD models.

The paper is concerned with the integration of building performance simulation within a collaborative/ multidisciplinary higher-education environment. The paper presents a semester-long setup in which a course attended by both architecture and engineering students and jointly taught by an architect and an engineer ultimately collaborate with an undergraduate architecture design studio on proposing upgrades to an existing building.

A wide range of scientifically validated Building Performance Simulation tools BPS is available internationally. The users of those tools are mainly researchers, physicists and experts who value empirical validation, analytical verification and calibration of uncertainty as defined by e.g. BESTEST. However, literature and comparative surveys indicate that most architects who use BPS tools in design practice are much more concerned with the (1) Usability and Information Management (UIM) of interface and (2) the Integration of Intelligent design Knowledge-Base (IIKB).

The daylight factor is usually one of the first daylight performance measures that simulation newcomers calculate. Apart from the intrinsic limitations of the daylight factor as a meaningful daylighting performance metric, little work has been done in the past as to how accurate one can actually expect simulation novices to simulate the daylight factor compared to an expert modeler. This paper presents the comparison of daylight factor predictions from a ‘best practice’ model of an L-shaped perimeter classroom to a total of 69 novice/student models.

In the current context of energy crisis and with the debate on climate change, low-energy buildings are required.

A genetic optimization approach has been used for the design of an external shading device in an office with a window and different glass characteristics. The primary energy consumption for heating cooling and lighting have been minimized. Simulations have been performed using the energy code ESP-r and lighting simulation package Radiance, the optimization loop has been driven by the software tool modeFRONTIER. Different optimal geometries for a panel shading device have been found.