Applications of Symlets for denoising and load forecasting

The Symmetrical Wavelet (Symlet) is proposed as a basis function in a Multi-Resolution Analysis (MRA) using the Discrete Wavelet Transform (DWT) to analyze power load consumption signals. Load forecasting is an important function in a utility as it supports maintenance, marketing, investment and production planning. Present load forecasting techniques rely heavily on past load patterns. These load consumption signals however are by nature corrupted by non-stationary and non-gaussian noise processes of which no models exists. Wavelet analysis is proposed in this case to denoise and isolate load trends in the consumption patterns. A one-week sample data is analysed to demonstrate the potential and benefits of such a scheme.