Model for generation of synthetic year-round weather data for use in building cooling load simulation

A set of hour-by-hour weather data is required in detailed analysis of the energy consumption in a building for providing air-conditioning. Time series analysis and modelling method has been applied to formulate a mathematical model for generation of synthetic weather data for Hong Kong for that purpose. The model developed based on weather data records of Hong Kong was a trivariate, autoregressive model of order one. Synthetic weather data generated by using that model have correlations slightly less than those of the sample data but it was found that a model of a much higher order would be required if any better modelling is to be achieved. Nevertheless, the generated data are considered suitable for use as representative weather conditions in hour-by-hour simulation of building cooling load. This paper describes the data generation model developed, a comparison of the generated data against the sample data and a brief summary of diagnostic checks done on the model.