A statistical approach to the development of a typical meteorological year for Hong Kong

Abstract

Measured hourly weather data for the 16-year period from 1979 to 1994 were analysed. Statistics techniques were used to develop a typical meteorological year (TMY) for building energy simulation. We used the Finkelstein-Schafer (FS) and the Kolmogrov-Smirnov (KS) two-sample statistics to analyse and select representative weather data for the formulation of a TMYfor Hong Kong. Building energy computer simulations using DOE-2 were carried out. It was found that the monthly electricity consumption profile from the TMY was close to that from the 16-year long-term weather data. The root-mean-square error was 27.7MWh, 3.9% of the long-term mean monthly consumption. On an annual basis, TMY predicted 1.5% more consumption than the long-term.