Thermal analysis basically means using a manual calculation or computer program to mathematically model the interplay of thermal processes within a building. There are a wide range of mathematical models used for this purpose, all of which vary significantly in both ease of implementation and comprehensiveness. Whilst none of the available models fully represent all aspects of the real world situation, there are persuasive arguments that they do not need to:
Simulation is the process of developing a simplified model of a complex system and using the model to analyse and predict the behaviour of the original system. Why simulate? The key reasons are that real-life systems are often difficult or impossible to analyse in all their complexity, and it is usually unnecessary to do so anyway. By carefully extracting from the real system the elements relevant to the stated requirements and ignoring the relatively insignificant ones (which is not as easy as it sounds), it is generally possible to develop a model that can be used to predict the behaviour of the real system accurately.
Aburdene, M.F. Computer Simulation of Dynamic Systems
Computer modelling of thermal performance is a complex but increasingly necessary part of modern building design. Up until recently the building design process has been sequential; first the building is designed by the architects and then the engineers size and fit the mechanical systems.
The architects rarely consider the dynamic thermal response of the building and the engineers rarely question the wisdom of the architects, ensuring that they over-design absolutely everything to protect their indemnity insurance. Thermal comfort requirements are usually reduced to a simple air temperature, neglecting a whole range of thermo-physiological environmental parameters like radiant temperature, air velocity and adaptive response.
Thankfully informed clients and progressive design firms are increasingly including engineers and environmental consultants as part of the initial design team. Many architectural practices are also developing thermal modelling capabilities in-house. It is often surprising to many architects involved in this new approach just how non-intuitive some of the heat transfer processes within a building can actually be. Many thermal modelling jobs turn up a host of interesting thermal effects that the designer would never have thought of but, once highlighted, become absurdly obvious.
Just like computer programs for weather prediction, the complexity and interplay of the actual mechanisms involved is far too complex for current computer systems and well beyond our own current understanding. However, we do know enough for modelling techniques to get it pretty close most of the time and to certainly pick up many things a rule-of-thumb or hand-calculation never would.
However, it is also important to understand the benefits and limitations of the various models as each one has a set of characteristics and idiosyncrasies that make it more or less suitable to certain conditions over others.
In broad terms most building thermal models fall into one of the following five major categories.
Steady state calculation methods are typically those that use U-values for the building fabric to calculate an overall building loss coefficient, and then multiply that by monthly or annual degree-day figures to estimate total heat losses/gains. Such methods were designed for manual calculation and have no mechanism for the inclusion of changes in direct solar gains, casual gains, long-wave radiation exchange, plant operation or the like.
Their major failing is an inability to factor in the dynamic effects of thermal mass in climates with a high diurnal range. As a result, these methods are being replaced by dynamic theories and play a diminishing role even at early design stages as desktop computing power becomes increasingly available.
In recent years a number of simplified methods of energy assessment have been produced which go some way to addressing dynamic performance. These methods are mostly based on regression techniques applied to the results of multiple parametric runs from a more powerful modelling systems. The results that emerge can often be reduced to simple relationships that can be represented in tabular or graphical form. Such methods are mostly seen in building advice schemes and in software that promotes a particular manufacturer or retailer's products.
By careful analysis and specification of boundary conditions, it is possible to solve the partial differential heat equations that govern the flow of heat within the building fabric. This sets up a series of equations that define the response of each surface to changes in temperature. Solving these incrementally as conditions change provides a means of modelling the dynamic response of a building. Two main branches of this method exist, time-domain response function and frequency-domain response function methods.
Time domain methods use recorded hourly weather data as input and calculate hourly internal conditions as output. Frequency domain methods are more concerned with the cyclical response of the system, exciting it with a known signal (usually a sin wave) and testing for harmonics and resonances within its response. Such methods are characterised by the use of a number of surface factors for each material that characterise its dynamic response in either domain. The most commonly used variant of this method is therefore known as the Response Factor method.
Another variant of this is the Admittance Method advocated by the UK Chartered Institute of Building Services Engineers. It uses a material characteristic known as the admittance of a surface (which is essentially a dynamic U-value), as well as thermal lag and decrement factors, to define their dynamic response. It is a simplified method in that it assumes a sinusoidal pattern of variation and calculates the instantaneous variation of internal conditions about their mean.
With the advent of powerful computing systems, many problems of varying complexity can be solved by shear brute force - calculating the heat flows between each layer of each material over the whole building fabric in relatively small time increments. These methods are characterised by systems that subdivide either each material into multiple equidistant layers or surfaces into multiple smaller segments of a 3D grid. The two main numerical techniques are the finite difference and finite element methods. Finite difference methods are most commonly applied to the problem of building energy modelling whilst finite element techniques are more often used in industrial and smaller scale processes.
The analogy that exists between electrical flow and heat flow has led to the construction of electrical analogue devices for the study of complex heat flow phenomena. This technique is useful as a research tool, allowing long-term simulations to be completed in a short elapsed time, but they do not currently see much application in the building design context.
There are a wide range of thermal modelling tools available. The following is by no means an exhaustive list, so if you know of a tool that you think is missing from this list, please e-mail us some details and it will be added willingly.
ESP-r is a dynamic thermal simulation environment for the analysis of energy and mass flows and environmental control systems within the built environment. ESP-r (the r stands for Research and EU Reference) allows researchers and designers to assess the manner in which actual weather patterns, occupant interactions, design parameter changes and control systems affect energy requirements and environmental states. It was developed by the Energy Systems Research Unit (ESRU), based within the Department of Mechanical Engineering at the University of Strathclyde in Glasgow, Scotland.
EnergyPlus is the merger of two older and well established tools, DOE2 and Blast. Basically it performs hourly simulations of buildings, air handling systems and central plant equipment to provide mechanical, environmental and architectural engineers with accurate estimates of a building's energy needs. The zone models of BLAST (Building Loads Analysis and System Thermodynamics), which are based on the fundamental heat balance method, are the industry standard for heating and cooling load calculations. BLAST output may be utilised in conjunction with the LCCID (Life Cycle Cost in Design) program to perform an economic analysis of the building/system/plant design. Developed by the US Department of Energy.
Building Design Advisor
BDA is a computer program that supports the concurrent, integrated use of multiple simulation tools and databases, through a single, object-based representation of building components and systems. Based on a comprehensive design theory, the BDA acts as a data manager and process controller, allowing building designers to benefit from the capabilities of multiple analysis and visualisation tools throughout the building design process. The BDA has a simple graphical user interface that is based on two main elements, the building browser and the decision desktop. Developed by Konstantinos Papamichael at Lawrence Berkeley National Laboratory.
Tas is a software package for the thermal analysis of buildings. It includes a 3D modeller, a thermal/energy analysis module, a systems/controls simulator and a 2D CFD package. There are also CAD links into the 3D modeller as well as report generation facilities. It is a complete solution for the thermal simulation of a building, and a powerful design tool in the optimisation of a buildings environmental, energy and comfort performance. Tas was developed by EDSL.
The NatHERS software was originally developed by the CSIRO, and is now administered as an initiative of the Ministerial Council on Energy. It is intended to provide quick, comprehensive and effective assessment of house design in an easy to use format, providing a rating out of 5 stars, where the greater number of stars represents a greater level of thermal comfort. It supports the state-based energy rating schemes as required under the Building Code of Australia regulated by the ABCB, and is assessed by ABSA accredited assessors. The NatHERS thermal calculation engine has now been superceded by the AccuRATE engine (see next), and is expected to be phased out completely by the end of November 2007.
BUNYIP is a computerised design procedure for accurately estimating the costs and amounts of energy used in commercial buildings. The software package, developed by the CSIRO Division of Building, Construction and Engineering, enables design office assessments of new buildings and retrofit proposals. It is especially useful in comparing design options and life-cycle costs. It can also be used for research and educational purposes. The core of BUNYIP is a component-based thermal model of a building and its services systems. This is backed up by a climatic database for numerous locations in Australia, New Zealand and South East Asia. The model allows almost any building to be analysed, as the user can select and combine building or system elements as desired. A large range of air handling, heating and cooling component models are included. Bunyip is no longer available.
ESPII is a suite of five computer programs used for estimating the energy consumption of a building over a given period of time taking into account the site location, the building structure and the type of building services installed to maintain the desired environmental conditions. It enables a designer to investigate many alternatives and make energy comparisons quickly and effectively for a very wide range of building configurations and air conditioning systems using actual measured climatic data. ESPII is a metric version of the American program ESP-II developed for APEC by Ferreira and Kalasinsky Associates Inc., New Bedford, Massachusetts, under the direction of the APEC Energy Analysis Committee. ACADS-BSG has the licensing rights of this program in Australia and New Zealand.
ECOTECT is a conceptual design analysis tool that features overshadowing, shading design, lighting, acoustic and wind analysis functions as well as thermal. It uses the CIBSE Admittance Method to calculate heating and cooling loads for any number of zones within a model. These loads factor in direct and indirect solar gains (with the option of calculating full overshadowing percentages for each hour of the day throughout the year), internal gains, inter-zonal heat flow and pull-down loads due to intermittent usage. It can display hourly internal temperatures and load breakdowns as well as annual temperature distributions and the effects of thermal mass. ECOTECT was developed by Dr. Andrew Marsh of SQUARE ONE research. Visit the ECOTECT home page here].
- The Building as an Integrated Dynamic System
- Jan L. M. Hensen, 1991.
- Computer Simulation of Dynamic Systems,
- Aburdene, M.F. Computer Simulation of Dynamic Systems
- C. Brown Publishers, Dubuque, IA, 1988.
- ESP-r On-line Course Notes
- DOE Tools directory