| Summary: | The liberalization and deregulation of electric markets forced the various participants to accommodate several challenges, including: a considerable accumulation of new generation capacity from renewable sources (fundamentally wind energy), the unpredictability associated with these new forms of generation and new consumption patterns, contributing to further electricity prices volatility (e.g. the Iberian market). Given the competitive framework in which market participants operate, the existence of efficient computational forecasting techniques is a distinctive factor. Based on these forecasts a suitable bidding strategy and an effective generation systems operation planning is achieved, together with an improved installed transmission capacity exploitation, results in maximized profits, all this contributing to a better energy resources utilization. This dissertation presents a new hybrid method for load and electricity prices forecasting, for one day ahead time horizon. The optimization scheme presented in this method, combines the efforts from different techniques, notably artificial neural networks, several optimization algorithms and wavelet transform. The method’s validation was made using different real case studies. The subsequent comparison (accuracy wise) with published results, in reference journals, validated the proposed hybrid method suitability.
|
|---|