SYNOPSISTo reduce weather data computer storage without loss of accuracy Discrete Fourier Series are often employed for representation of hourly Solar Radiation inputs. This paper applies a method [ I ] which is based on the circular convolution theorem for determining the Fourier coefficients (G,) of the solar radiation transmittance G(t) through glazings from the coefficients t, and I , of the discrete Fourier Series representations of the tranmittance function t(t) and of the coefficient I(t), respectively.As shown, 33 harmonics may represent the whole series of hourly data from one week, which means a computer file size reduction of at least 80%. The results of transmitted hourly radiation for South and + 25" South-West (SW) oriented windows are in true agreement with the ones obtained by another procedure [2] based on computer simulation.
INTRODUCTIONTo represent weather inputs in solar energy analysis and building thermal applications, the Fourier Series are widely used [3, 41. Among the several inputs, the one with the most bearing is the incident solar radiation. The portion of the incident radiation which is transmitted through windows or glazing collectors must also be taken into account.In a former paper [2] we presented a method of computer simulation to evaluate the solar radiation transmitted through windows every hour for whatever orientation and the amount absorbed after reflection on the walls of the building. The procedure described is based on solar geometric relations and uses the transmission coefficient of the glass and the absorbtance coefficient of the walls.To evaluate the transmitted solar radiation for the period of a whole year a large amount of computer storage is needed. The use of the Discrete Fourier Transform (DFT) reduces computational requirement substantially since it avoids the storage of the whole series in the time domain.