A lot of work in the field of thermal modelling of buildings has been done using Fourier analysis by representing a set of data of time dependent variables by a Fourier Series. Major researchers have used a system of Continuous Fourier Transform, i.e. approximating a set of sampled data to be a continuous function which is periodic in nature but linear between two sampled instances and calculating the Fourier coefficients by using integral function and solving for six harmonics. This method, apart from being approximate, has mathematical limitations due to the presence of the integral. The calculations are complicated and time consuming. The approach of Discrete Fourier Transform, commonly used in digital signal processing, is introduced in this paper as a better alternative for solar thermal modelling, in that it provides a simpler approach, with as accurate results as the continuous method when six harmonics are considered and exact results at the sampled instances when Nyquist's criterion is met. This method takes only 1/120th of the computing time required for the continuous method.
展开▼
