you are here > DIT Dublin Energy Lab > Research Themes > zero emissons building > numerical methods for building energy simulation

dublin energy lab,
focas institute,
dublin institute of technology,
dublin 8,

t/ +353 1 402 7912
f/ +353 1 402 7904



investigation and development of numerical methods for building energy simulation

A large fraction of total energy use is associated with moderation of environmental conditions within buildings. The cost of this energy and the associated CO2 emissions make energy conscious design, operation and refurbishment of buildings imperative. To this end building energy simulation tools are increasingly used to determine peak heating and cooling loads, size thermal plant, anticipate annual energy consumption and analyze thermal comfort.

A building energy model comprises a large set of differential equations originating in sub-models which describe heat transfer modes at building elements. Implicit numerical solution methods are suggested by the stiffness of the problem and the better-known simulation packages utilise them. However, the more commonly used methods are drawn from a limited, ‘traditional’ set and are often just marginally A-stable. Stronger stability properties are desirable to prevent persistent, unrealistic oscillations in the solution, and more efficient solution methods would offer improved accuracy which could, if desired, be traded for greater speed of execution.

Test programs, which include efficient error estimation and automatic time-step adjustment, are written for established and recently developed numerical methods. A comprehensive and discerning test methodology is applied and a number of A-stable and L-stable methods are identified or developed which offer superior stability and/or greater computational efficiency when compared with those currently in use. For example, a proposed development of the solver within the European reference model, ESP-r, is shown to produce about 30% less error than the currently used method. It is worth noting that improvements in the numerical solver used in conjunction with a building energy model have a global impact on performance whereas enhancement of any of the constituent models probably has a more limited effect.


DIT - School of Civil & Building Services Engineering

Dr Michael Crowley


Prof. M.S.J. Hashmi, Dublin City University