COMPRESSION SOLAR RADIATION DATA USING HAAR AND DAUBECHIES WAVELETS

In this paper a study was carried out to use the Haar and Daubechies wavelets for solar radiation data compression. The focus is on solar radiation data that was collected at Universiti Teknologi PETRONAS. The data was collected by using the state-of-art pyranometers mounted on an advance sun tracking system. The measuring system that is connected to a computerised data logging system can record the direct and diffuse solar radiation data. The resolution of the measurements can be set, but the equipment has the ability to measure data for every second. This allows for an effective and comprehensive numerical analysis to be carried out both using the wavelet approach as well as the conventional statistical analysis. The results obtained in the analysis show that the longer the filters of wavelets (in our case Daubechies), the better the compression results that can be achieved. Matlab is used to represent all the numerical results. The highly transient nature of solar radiation data is usually averaged out by using different statistical approaches. Usually there are attempts to incorporate probabilistic approaches, but have been proven to be futile, as there is no single equation that can be used for forecasting purposes. So therefore the trend analysis must be based on compressed historical data, and location sensitive time series equation must be developed as well.