8 – Dynamic Electric Load Forecasting

摘要: Publisher Summary This chapter presents approaches for long-term and mid-term electric power load forecasting. It implements strong short-term correlations of daily (24 hours) yearly (52 weeks) behavior to predict future demand. forecasts weekly average profiles 24 hours a day with lead time from several weeks few years. uses simple linear regression models previous data augmented annual growth introduces the Kalman filtering algorithm moving window weather, both crisp fuzzy loads, estimate optimal forecast parameters. recursive least error squares estimation, as dynamic load-forecast Most load-forecasting methods are dedicated forecasting, but not many or intermediate-term The great importance forecasting utility planning its economic consequences is encouraging development in research improve accuracy. By nature, complex problem. Among other factors, accuracy extremely influenced by weather well social community that load. These factors difficult horizon. Conversely, although affected habits, small enough high